diff --git a/CLAUDE.md b/CLAUDE.md
index f5d5cf5..745d561 100644
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -99,12 +99,14 @@ Semcode uses the following search order to locate the `.semcode.db` database dir
 
 **For semcode-index:**
 1. **-d flag**: If provided, use the specified path (direct database path or parent directory containing `.semcode.db`)
-2. **Source directory**: Look for `.semcode.db` in the source directory specified by `-s`
-3. **Current directory**: Fall back to `./.semcode.db` in the current working directory
+2. **SEMCODE_DB environment variable**: Same path semantics as `-d`
+3. **Source directory**: Look for `.semcode.db` in the source directory specified by `-s`
+4. **Current directory**: Fall back to `./.semcode.db` in the current working directory
 
 **For semcode (query tool), semcode-mcp, and semcode-lsp:**
 1. **-d flag / configuration**: If provided, use the specified path (direct database path or parent directory containing `.semcode.db`)
-2. **Workspace/Current directory**: Use `./.semcode.db` in the workspace or current working directory
+2. **SEMCODE_DB environment variable**: Same path semantics as `-d`
+3. **Workspace/Current directory**: Use `./.semcode.db` in the workspace or current working directory
 
 The `-d` flag can specify either:
 - A direct path to the database directory (e.g., `./my-custom.db`)
diff --git a/Cargo.lock b/Cargo.lock
index ff77c04..c8993b1 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -98,7 +98,7 @@ version = "1.1.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "40c48f72fd53cd289104fc64099abca73db4166ad86ea0b4341abe65af83dadc"
 dependencies = [
- "windows-sys 0.60.2",
+ "windows-sys 0.61.2",
 ]
 
 [[package]]
@@ -109,7 +109,7 @@ checksum = "291e6a250ff86cd4a820112fb8898808a366d8f9f58ce16d1f538353ad55747d"
 dependencies = [
  "anstyle",
  "once_cell_polyfill",
- "windows-sys 0.60.2",
+ "windows-sys 0.61.2",
 ]
 
 [[package]]
@@ -120,9 +120,9 @@ checksum = "7f202df86484c868dbad7eaa557ef785d5c66295e41b460ef922eca0723b842c"
 
 [[package]]
 name = "arc-swap"
-version = "1.9.0"
+version = "1.9.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a07d1f37ff60921c83bdfc7407723bdefe89b44b98a9b772f225c8f9d67141a6"
+checksum = "6a3a1fd6f75306b68087b831f025c712524bcb19aad54e557b1129cfa0a2b207"
 dependencies = [
  "rustversion",
 ]
@@ -284,7 +284,7 @@ dependencies = [
  "arrow-schema",
  "chrono",
  "half",
- "indexmap 2.13.0",
+ "indexmap 2.14.0",
  "itoa",
  "lexical-core",
  "memchr",
@@ -327,7 +327,7 @@ version = "57.3.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "8c872d36b7bf2a6a6a2b40de9156265f0242910791db366a2c17476ba8330d68"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "serde_core",
  "serde_json",
 ]
@@ -377,9 +377,9 @@ dependencies = [
 
 [[package]]
 name = "async-compression"
-version = "0.4.41"
+version = "0.4.42"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d0f9ee0f6e02ffd7ad5816e9464499fba7b3effd01123b515c41d1697c43dad1"
+checksum = "e79b3f8a79cccc2898f31920fc69f304859b3bd567490f75ebf51ae1c792a9ac"
 dependencies = [
  "compression-codecs",
  "compression-core",
@@ -444,17 +444,6 @@ version = "1.1.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "1505bd5d3d116872e7271a6d4e16d81d0c8570876c8de68093a09ac269d8aac0"
 
-[[package]]
-name = "auto_impl"
-version = "1.3.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ffdcb70bdbc4d478427380519163274ac86e52916e10f0a8889adf0f96d3fee7"
-dependencies = [
- "proc-macro2",
- "quote",
- "syn 2.0.117",
-]
-
 [[package]]
 name = "autocfg"
 version = "1.5.0"
@@ -463,9 +452,9 @@ checksum = "c08606f8c3cbf4ce6ec8e28fb0014a2c086708fe954eaa885384a6165172e7e8"
 
 [[package]]
 name = "aws-lc-rs"
-version = "1.16.2"
+version = "1.16.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a054912289d18629dc78375ba2c3726a3afe3ff71b4edba9dedfca0e3446d1fc"
+checksum = "0ec6fb3fe69024a75fa7e1bfb48aa6cf59706a101658ea01bfd33b2b248a038f"
 dependencies = [
  "aws-lc-sys",
  "zeroize",
@@ -473,9 +462,9 @@ dependencies = [
 
 [[package]]
 name = "aws-lc-sys"
-version = "0.39.0"
+version = "0.40.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1fa7e52a4c5c547c741610a2c6f123f3881e409b714cd27e6798ef020c514f0a"
+checksum = "f50037ee5e1e41e7b8f9d161680a725bd1626cb6f8c7e901f91f942850852fe7"
 dependencies = [
  "cc",
  "cmake",
@@ -485,9 +474,9 @@ dependencies = [
 
 [[package]]
 name = "axum"
-version = "0.8.8"
+version = "0.8.9"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8b52af3cb4058c895d37317bb27508dccc8e5f2d39454016b297bf4a400597b8"
+checksum = "31b698c5f9a010f6573133b09e0de5408834d0c82f8d7475a89fc1867a71cd90"
 dependencies = [
  "axum-core",
  "base64 0.22.1",
@@ -512,8 +501,8 @@ dependencies = [
  "sha1",
  "sync_wrapper",
  "tokio",
- "tokio-tungstenite 0.28.0",
- "tower 0.5.3",
+ "tokio-tungstenite",
+ "tower",
  "tower-layer",
  "tower-service",
  "tracing",
@@ -565,15 +554,9 @@ dependencies = [
 
 [[package]]
 name = "bitflags"
-version = "1.3.2"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bef38d45163c2f1dde094a7dfd33ccf595c92905c8f8f4fdc18d06fb1037718a"
-
-[[package]]
-name = "bitflags"
-version = "2.11.0"
+version = "2.11.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "843867be96c8daad0d758b57df9392b6d8d271134fce549de6ce169ff98a92af"
+checksum = "c4512299f36f043ab09a583e57bceb5a5aab7a73db1805848e8fef3c9e8c78b3"
 
 [[package]]
 name = "bitpacking"
@@ -607,16 +590,16 @@ dependencies = [
 
 [[package]]
 name = "blake3"
-version = "1.8.3"
+version = "1.8.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2468ef7d57b3fb7e16b576e8377cdbde2320c60e1491e961d11da40fc4f02a2d"
+checksum = "0aa83c34e62843d924f905e0f5c866eb1dd6545fc4d719e803d9ba6030371fce"
 dependencies = [
  "arrayref",
  "arrayvec",
  "cc",
  "cfg-if",
  "constant_time_eq",
- "cpufeatures",
+ "cpufeatures 0.3.0",
 ]
 
 [[package]]
@@ -653,6 +636,12 @@ dependencies = [
  "syn 2.0.117",
 ]
 
+[[package]]
+name = "borrow-or-share"
+version = "0.2.4"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "dc0b364ead1874514c8c2855ab558056ebfeb775653e7ae45ff72f28f8f3166c"
+
 [[package]]
 name = "brotli"
 version = "8.0.2"
@@ -726,9 +715,9 @@ dependencies = [
 
 [[package]]
 name = "cc"
-version = "1.2.57"
+version = "1.2.61"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7a0dd1ca384932ff3641c8718a02769f1698e7563dc6974ffd03346116310423"
+checksum = "d16d90359e986641506914ba71350897565610e87ce0ad9e6f28569db3dd5c6d"
 dependencies = [
  "find-msvc-tools",
  "jobserver",
@@ -742,12 +731,6 @@ version = "0.4.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "4f4c707c6a209cbe82d10abd08e1ea8995e9ea937d2550646e02798948992be0"
 
-[[package]]
-name = "cesu8"
-version = "1.1.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6d43a04d8753f35258c91f8ec639f792891f748a1edbd759cf1dcea3382ad83c"
-
 [[package]]
 name = "cfg-if"
 version = "1.0.4"
@@ -796,9 +779,9 @@ dependencies = [
 
 [[package]]
 name = "clap"
-version = "4.6.0"
+version = "4.6.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b193af5b67834b676abd72466a96c1024e6a6ad978a1f484bd90b85c94041351"
+checksum = "1ddb117e43bbf7dacf0a4190fef4d345b9bad68dfc649cb349e7d17d28428e51"
 dependencies = [
  "clap_builder",
  "clap_derive",
@@ -818,9 +801,9 @@ dependencies = [
 
 [[package]]
 name = "clap_derive"
-version = "4.6.0"
+version = "4.6.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1110bd8a634a1ab8cb04345d8d878267d57c3cf1b38d91b71af6686408bbca6a"
+checksum = "f2ce8604710f6733aa641a2b3731eaa1e8b3d9973d5e3565da11800813f997a9"
 dependencies = [
  "heck",
  "proc-macro2",
@@ -854,9 +837,9 @@ dependencies = [
 
 [[package]]
 name = "cmake"
-version = "0.1.57"
+version = "0.1.58"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "75443c44cd6b379beb8c5b45d85d0773baf31cce901fe7bb252f4eff3008ef7d"
+checksum = "c0f78a02292a74a88ac736019ab962ece0bc380e3f977bf72e376c5d78ff0678"
 dependencies = [
  "cc",
 ]
@@ -873,7 +856,7 @@ version = "3.1.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "faf9468729b8cbcea668e36183cb69d317348c2e08e994829fb56ebfdfbaac34"
 dependencies = [
- "windows-sys 0.52.0",
+ "windows-sys 0.61.2",
 ]
 
 [[package]]
@@ -913,9 +896,9 @@ dependencies = [
 
 [[package]]
 name = "compression-codecs"
-version = "0.4.37"
+version = "0.4.38"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "eb7b51a7d9c967fc26773061ba86150f19c50c0d65c887cb1fbe295fd16619b7"
+checksum = "ce2548391e9c1929c21bf6aa2680af86fe4c1b33e6cea9ac1cfeec0bd11218cf"
 dependencies = [
  "compression-core",
  "flate2",
@@ -924,9 +907,9 @@ dependencies = [
 
 [[package]]
 name = "compression-core"
-version = "0.4.31"
+version = "0.4.32"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "75984efb6ed102a0d42db99afb6c1948f0380d1d91808d5529916e6c08b49d8d"
+checksum = "cc14f565cf027a105f7a44ccf9e5b424348421a1d8952a8fc9d499d313107789"
 
 [[package]]
 name = "concurrent-queue"
@@ -1023,6 +1006,15 @@ dependencies = [
  "libc",
 ]
 
+[[package]]
+name = "cpufeatures"
+version = "0.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "8b2a41393f66f16b0823bb79094d54ac5fbd34ab292ddafb9a0456ac9f87d201"
+dependencies = [
+ "libc",
+]
+
 [[package]]
 name = "crc32fast"
 version = "1.5.0"
@@ -1139,9 +1131,9 @@ dependencies = [
 
 [[package]]
 name = "curl-sys"
-version = "0.4.86+curl-8.19.0"
+version = "0.4.88+curl-8.20.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3a1dd6a487cf4532ce0d801634b82aa2deb7c9c3ed930b9dadfce904df000745"
+checksum = "644816de6547255eff4e491a1dda1c19b7237f00b62a61e6e64859ce4f2906d0"
 dependencies = [
  "cc",
  "libc",
@@ -1149,7 +1141,7 @@ dependencies = [
  "openssl-sys",
  "pkg-config",
  "vcpkg",
- "windows-sys 0.59.0",
+ "windows-sys 0.61.2",
 ]
 
 [[package]]
@@ -1223,26 +1215,13 @@ dependencies = [
 
 [[package]]
 name = "dary_heap"
-version = "0.3.8"
+version = "0.3.9"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "06d2e3287df1c007e74221c49ca10a95d557349e54b3a75dc2fb14712c751f04"
+checksum = "8b1e3a325bc115f096c8b77bbf027a7c2592230e70be2d985be950d3d5e60ebe"
 dependencies = [
  "serde",
 ]
 
-[[package]]
-name = "dashmap"
-version = "5.5.3"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "978747c1d849a7d2ee5e8adc0159961c48fb7e5db2f06af6723b80123bb53856"
-dependencies = [
- "cfg-if",
- "hashbrown 0.14.5",
- "lock_api",
- "once_cell",
- "parking_lot_core",
-]
-
 [[package]]
 name = "dashmap"
 version = "6.1.0"
@@ -1259,15 +1238,15 @@ dependencies = [
 
 [[package]]
 name = "data-encoding"
-version = "2.10.0"
+version = "2.11.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d7a1e2f27636f116493b8b860f5546edb47c8d8f8ea73e1d2a20be88e28d1fea"
+checksum = "a4ae5f15dda3c708c0ade84bfee31ccab44a3da4f88015ed22f63732abe300c8"
 
 [[package]]
 name = "datafusion"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "43c18ba387f9c05ac1f3be32a73f8f3cc6c1cfc43e5d4b7a8e5b0d3a5eb48dc7"
+checksum = "7541353e77dc7262b71ca27be07d8393661737e3a73b5d1b1c6f7d814c64fa2a"
 dependencies = [
  "arrow",
  "arrow-schema",
@@ -1303,7 +1282,7 @@ dependencies = [
  "log",
  "object_store",
  "parking_lot",
- "rand 0.9.2",
+ "rand 0.9.4",
  "regex",
  "sqlparser",
  "tempfile",
@@ -1314,13 +1293,13 @@ dependencies = [
 
 [[package]]
 name = "datafusion-catalog"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3c75a4ce672b27fb8423810efb92a3600027717a1664d06a2c307eeeabcec694"
+checksum = "9997731f90fa5398ef831ad0e69600f92c861b79c0d38bd1a29b6f0e3a0ce4c8"
 dependencies = [
  "arrow",
  "async-trait",
- "dashmap 6.1.0",
+ "dashmap",
  "datafusion-common",
  "datafusion-common-runtime",
  "datafusion-datasource",
@@ -1339,9 +1318,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-catalog-listing"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2c8b9a3795ffb46bf4957a34c67d89a67558b311ae455c8d4295ff2115eeea50"
+checksum = "2b30a3dd50dec860c9559275c8d97d9de602e611237a6ecfbda0b3b63b872352"
 dependencies = [
  "arrow",
  "async-trait",
@@ -1362,9 +1341,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-common"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "205dc1e20441973f470e6b7ef87626a3b9187970e5106058fef1b713047f770c"
+checksum = "d551054acec0398ca604512310b77ce05c46f66e54b54d48200a686e385cca4e"
 dependencies = [
  "ahash",
  "arrow",
@@ -1372,7 +1351,7 @@ dependencies = [
  "chrono",
  "half",
  "hashbrown 0.16.1",
- "indexmap 2.13.0",
+ "indexmap 2.14.0",
  "libc",
  "log",
  "object_store",
@@ -1384,9 +1363,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-common-runtime"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8cf5880c02ff6f5f11fb5bc19211789fb32fd3c53d79b7d6cb2b12e401312ba0"
+checksum = "567d40e285f5b79f8737b576605721cd6c1133b5d2b00bdbd5d9838d90d0812f"
 dependencies = [
  "futures",
  "log",
@@ -1395,9 +1374,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-datasource"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bc614d6e709450e29b7b032a42c1bdb705f166a6b2edef7bed7c7897eb905499"
+checksum = "27d2668f51b3b30befae2207472569e37807fdedd1d14da58acc6f8ca6257eae"
 dependencies = [
  "arrow",
  "async-trait",
@@ -1417,16 +1396,16 @@ dependencies = [
  "itertools 0.14.0",
  "log",
  "object_store",
- "rand 0.9.2",
+ "rand 0.9.4",
  "tokio",
  "url",
 ]
 
 [[package]]
 name = "datafusion-datasource-arrow"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6e497d5fc48dac7ce86f6b4fb09a3a494385774af301ff20ec91aebfae9b05b4"
+checksum = "e02e1b3e3a8ec55f1f62de4252b0407c8567363d056078769a197e24fc834a0f"
 dependencies = [
  "arrow",
  "arrow-ipc",
@@ -1448,9 +1427,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-datasource-csv"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0dfc250cad940d0327ca2e9109dc98830892d17a3d6b2ca11d68570e872cf379"
+checksum = "b559d7bf87d4f900f847baba8509634f838d9718695389e903604cdcccdb01f3"
 dependencies = [
  "arrow",
  "async-trait",
@@ -1471,9 +1450,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-datasource-json"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c91e9677ed62833b0e8129dec0d1a8f3c9bb7590bd6dd714a43e4c3b663e4aa0"
+checksum = "250e2d7591ba8b638f063854650faa40bca4e8bd4059b2ece8836f6388d02db4"
 dependencies = [
  "arrow",
  "async-trait",
@@ -1493,36 +1472,36 @@ dependencies = [
 
 [[package]]
 name = "datafusion-doc"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3e13e5fe3447baa0584b61ee8644086e007e1ef6e58f4be48bc8a72417854729"
+checksum = "b9496cb0db222dbb9a3735760ceca7fc56f35e1d5502c38d0caa77a81e9c1f6a"
 
 [[package]]
 name = "datafusion-execution"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "48a6cc03e34899a54546b229235f7b192634c8e832f78a267f0989b18216c56d"
+checksum = "dc45d23c516ed8d3637751e44e09e21b45b3f58b473c802dddd1f1ad4fe435ff"
 dependencies = [
  "arrow",
  "async-trait",
  "chrono",
- "dashmap 6.1.0",
+ "dashmap",
  "datafusion-common",
  "datafusion-expr",
  "futures",
  "log",
  "object_store",
  "parking_lot",
- "rand 0.9.2",
+ "rand 0.9.4",
  "tempfile",
  "url",
 ]
 
 [[package]]
 name = "datafusion-expr"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ee3315d87eca7a7df58e52a1fb43b4c4171b545fd30ffc3102945c162a9f6ddb"
+checksum = "63dd30526d2db4fda6440806a41e4676334a94bc0596cc9cc2a0efed20ef2c44"
 dependencies = [
  "arrow",
  "async-trait",
@@ -1533,7 +1512,7 @@ dependencies = [
  "datafusion-functions-aggregate-common",
  "datafusion-functions-window-common",
  "datafusion-physical-expr-common",
- "indexmap 2.13.0",
+ "indexmap 2.14.0",
  "itertools 0.14.0",
  "paste",
  "serde_json",
@@ -1542,22 +1521,22 @@ dependencies = [
 
 [[package]]
 name = "datafusion-expr-common"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "98c6d83feae0753799f933a2c47dfd15980c6947960cb95ed60f5c1f885548b3"
+checksum = "1b486b5f6255d40976b88bb83813b0d035a8333e0ec39864824e78068cf42fa6"
 dependencies = [
  "arrow",
  "datafusion-common",
- "indexmap 2.13.0",
+ "indexmap 2.14.0",
  "itertools 0.14.0",
  "paste",
 ]
 
 [[package]]
 name = "datafusion-functions"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "49b82962015cc3db4d7662459c9f7fcda0591b5edacb8af1cf3bc3031f274800"
+checksum = "07356c94118d881130dd0ffbff127540407d969c8978736e324edcd6c41cd48f"
 dependencies = [
  "arrow",
  "arrow-buffer",
@@ -1577,7 +1556,7 @@ dependencies = [
  "log",
  "md-5",
  "num-traits",
- "rand 0.9.2",
+ "rand 0.9.4",
  "regex",
  "sha2",
  "unicode-segmentation",
@@ -1586,9 +1565,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-functions-aggregate"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4e42c227d9e55a6c8041785d4a8a117e4de531033d480aae10984247ac62e27e"
+checksum = "b644f9cf696df9233ce6958b9807666d78563b56f923267474dd6c07795f1f8f"
 dependencies = [
  "ahash",
  "arrow",
@@ -1607,9 +1586,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-functions-aggregate-common"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "cead3cfed825b0b688700f4338d281cd7857e4907775a5b9554c083edd5f3f95"
+checksum = "c1de2deaaabe8923ce9ea9f29c47bbb4ee14f67ea2fe1ab5398d9bbebcf86e56"
 dependencies = [
  "ahash",
  "arrow",
@@ -1620,9 +1599,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-functions-nested"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "62ea99612970aebab8cf864d02eb3d296bbab7f4881e1023d282b57fe431b201"
+checksum = "552f8d92e4331ee91d23c02d12bb6acf32cbfd5215117e01c0fb63cd4b15af1a"
 dependencies = [
  "arrow",
  "arrow-ord",
@@ -1643,9 +1622,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-functions-table"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d83dbf3ab8b9af6f209b068825a7adbd3b88bf276f2a1ec14ba09567b97f5674"
+checksum = "970fd0cdd3df8802b9a9975ff600998289ba9d46682a4f7285cba4820c9ada78"
 dependencies = [
  "arrow",
  "async-trait",
@@ -1659,9 +1638,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-functions-window"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "732edabe07496e2fc5a1e57a284d7a36edcea445a2821119770a0dea624b472c"
+checksum = "40b4c21a7c8a986a1866c0a87ab756d0bbf7b5f41f306009fa2d9af79c52ed31"
 dependencies = [
  "arrow",
  "datafusion-common",
@@ -1677,9 +1656,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-functions-window-common"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e0c6e30e09700799bd52adce8c377ab03dda96e73a623e4803a31ad94fe7ce14"
+checksum = "b1210ad73b8b3211aeaf4a42bef9bd7a2b7fce3ec119a478831f18c6ff7f7b93"
 dependencies = [
  "datafusion-common",
  "datafusion-physical-expr-common",
@@ -1687,9 +1666,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-macros"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "402f2a8ed70fb99a18f71580a1fe338604222a3d32ddeac6e72c5b34feea2d4d"
+checksum = "aaa566a963013a38681ad82a727a654bc7feb19632426aea8c3412d415d200c5"
 dependencies = [
  "datafusion-doc",
  "quote",
@@ -1698,9 +1677,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-optimizer"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "99f32edb8ba12f08138f86c09b80fae3d4a320551262fa06b91d8a8cb3065a5b"
+checksum = "ff9aa82b240252a88dee118372f9b9757c545ab9e53c0736bebab2e7da0ef1f2"
 dependencies = [
  "arrow",
  "chrono",
@@ -1708,7 +1687,7 @@ dependencies = [
  "datafusion-expr",
  "datafusion-expr-common",
  "datafusion-physical-expr",
- "indexmap 2.13.0",
+ "indexmap 2.14.0",
  "itertools 0.14.0",
  "log",
  "regex",
@@ -1717,9 +1696,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-physical-expr"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "987c5e29e96186589301b42e25aa7d11bbe319a73eb02ef8d755edc55b5b89fc"
+checksum = "7d48022b8af9988c1d852644f9e8b5584c490659769a550c5e8d39457a1da0a5"
 dependencies = [
  "ahash",
  "arrow",
@@ -1730,7 +1709,7 @@ dependencies = [
  "datafusion-physical-expr-common",
  "half",
  "hashbrown 0.16.1",
- "indexmap 2.13.0",
+ "indexmap 2.14.0",
  "itertools 0.14.0",
  "parking_lot",
  "paste",
@@ -1740,9 +1719,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-physical-expr-adapter"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1de89d0afa08b6686697bd8a6bac4ba2cd44c7003356e1bce6114d5a93f94b5c"
+checksum = "ae7a8abc0b4fe624000972a9b145b30b7f1b680bffaa950ea53f78d9b21c27c3"
 dependencies = [
  "arrow",
  "datafusion-common",
@@ -1755,9 +1734,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-physical-expr-common"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "602d1970c0fe87f1c3a36665d131fbfe1c4379d35f8fc5ec43a362229ad2954d"
+checksum = "147253ca3e6b9d59c162de64c02800973018660e13340dd1886dd038d17ac429"
 dependencies = [
  "ahash",
  "arrow",
@@ -1765,16 +1744,16 @@ dependencies = [
  "datafusion-common",
  "datafusion-expr-common",
  "hashbrown 0.16.1",
- "indexmap 2.13.0",
+ "indexmap 2.14.0",
  "itertools 0.14.0",
  "parking_lot",
 ]
 
 [[package]]
 name = "datafusion-physical-optimizer"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b24d704b6385ebe27c756a12e5ba15684576d3b47aeca79cc9fb09480236dc32"
+checksum = "689156bb2282107b6239db8d7ef44b4dab10a9b33d3491a0c74acac5e4fedd72"
 dependencies = [
  "arrow",
  "datafusion-common",
@@ -1790,9 +1769,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-physical-plan"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c21d94141ea5043e98793f170798e9c1887095813b8291c5260599341e383a38"
+checksum = "68253dc0ee5330aa558b2549c9b0da5af9fc17d753ae73022939014ad616fc28"
 dependencies = [
  "ahash",
  "arrow",
@@ -1811,7 +1790,7 @@ dependencies = [
  "futures",
  "half",
  "hashbrown 0.16.1",
- "indexmap 2.13.0",
+ "indexmap 2.14.0",
  "itertools 0.14.0",
  "log",
  "parking_lot",
@@ -1821,9 +1800,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-pruning"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1a68cce43d18c0dfac95cacd74e70565f7e2fb12b9ed41e2d312f0fa837626b1"
+checksum = "0fcad240a54d0b1d3e8f668398900260a53122d522b2102ab57218590decacd6"
 dependencies = [
  "arrow",
  "datafusion-common",
@@ -1838,9 +1817,9 @@ dependencies = [
 
 [[package]]
 name = "datafusion-session"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6b4e1c40a0b1896aed4a4504145c2eb7fa9b9da13c2d04b40a4767a09f076199"
+checksum = "f58e83a68bb67007a8fcbf005c44cefe441270c7ee7f6dee10c0e0109b556f6d"
 dependencies = [
  "async-trait",
  "datafusion-common",
@@ -1852,16 +1831,16 @@ dependencies = [
 
 [[package]]
 name = "datafusion-sql"
-version = "52.4.0"
+version = "52.5.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2f1891e5b106d1d73c7fe403bd8a265d19c3977edc17f60808daf26c2fe65ffb"
+checksum = "be53e9eb55db0fbb8980bb6d87f2435b0524acf4c718ed54a57cabbb299b2ab3"
 dependencies = [
  "arrow",
  "bigdecimal",
  "chrono",
  "datafusion-common",
  "datafusion-expr",
- "indexmap 2.13.0",
+ "indexmap 2.14.0",
  "log",
  "regex",
  "sqlparser",
@@ -1957,7 +1936,7 @@ dependencies = [
  "libc",
  "option-ext",
  "redox_users",
- "windows-sys 0.60.2",
+ "windows-sys 0.61.2",
 ]
 
 [[package]]
@@ -2029,7 +2008,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "39cab71617ae0d63f51a36d69f866391735b51691dbda63cf6f96d042b63efeb"
 dependencies = [
  "libc",
- "windows-sys 0.52.0",
+ "windows-sys 0.61.2",
 ]
 
 [[package]]
@@ -2049,9 +2028,9 @@ dependencies = [
 
 [[package]]
 name = "ethnum"
-version = "1.5.2"
+version = "1.5.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ca81e6b4777c89fd810c25a4be2b1bd93ea034fbe58e6a75216a34c6b82c539b"
+checksum = "40404c3f5f511ec4da6fe866ddf6a717c309fdbb69fbbad7b0f3edab8f2e835f"
 
 [[package]]
 name = "event-listener"
@@ -2098,9 +2077,9 @@ dependencies = [
 
 [[package]]
 name = "fastrand"
-version = "2.3.0"
+version = "2.4.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "37909eebbb50d72f9059c3b6d82c0463f2ff062c9e95845c43a6c9c0355411be"
+checksum = "9f1f227452a390804cdb637b74a86990f2a7d7ba4b7d5693aac9b4dd6defd8d6"
 
 [[package]]
 name = "fd-lock"
@@ -2110,7 +2089,7 @@ checksum = "0ce92ff622d6dadf7349484f42c93271a0d49b7cc4d466a936405bacbe10aa78"
 dependencies = [
  "cfg-if",
  "rustix 1.1.4",
- "windows-sys 0.52.0",
+ "windows-sys 0.59.0",
 ]
 
 [[package]]
@@ -2142,7 +2121,7 @@ version = "25.12.19"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "35f6839d7b3b98adde531effaf34f0c2badc6f4735d26fe74709d8e513a96ef3"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "rustc_version",
 ]
 
@@ -2156,6 +2135,17 @@ dependencies = [
  "miniz_oxide",
 ]
 
+[[package]]
+name = "fluent-uri"
+version = "0.4.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "bc74ac4d8359ae70623506d512209619e5cf8f347124910440dbc221714b328e"
+dependencies = [
+ "borrow-or-share",
+ "ref-cast",
+ "serde",
+]
+
 [[package]]
 name = "fnv"
 version = "1.0.7"
@@ -2201,12 +2191,12 @@ checksum = "42703706b716c37f96a77aea830392ad231f44c9e9a67872fa5548707e11b11c"
 
 [[package]]
 name = "fsst"
-version = "3.0.1"
+version = "4.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a32ddfc5478379cd1782bdd9d7d1411063f563e5b338fc73bafe5916451a5b9d"
+checksum = "2195cc7f87e84bd695586137de99605e7e9579b26ec5e01b82960ddb4d0922f2"
 dependencies = [
  "arrow-array",
- "rand 0.9.2",
+ "rand 0.9.4",
 ]
 
 [[package]]
@@ -2438,7 +2428,7 @@ dependencies = [
  "regex",
  "signal-hook",
  "smallvec",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -2479,15 +2469,15 @@ dependencies = [
  "gix-trace",
  "kstring",
  "smallvec",
- "thiserror 2.0.18",
+ "thiserror",
  "unicode-bom",
 ]
 
 [[package]]
 name = "gix-bitmap"
-version = "0.3.0"
+version = "0.3.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e7add20f40d060db8c9b1314d499bac6ed7480f33eb113ce3e1cf5d6ff85d989"
+checksum = "1ecbfc77ec6852294e341ecc305a490b59f2813e6ca42d79efda5099dcab1894"
 dependencies = [
  "gix-error",
 ]
@@ -2509,23 +2499,23 @@ dependencies = [
  "gix-traverse",
  "gix-worktree",
  "smallvec",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
 name = "gix-chunk"
-version = "0.7.0"
+version = "0.7.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1096b6608fbe5d27fb4984e20f992b4e76fb8c613f6acb87d07c5831b53a6959"
+checksum = "edf288be9b60fe7231de03771faa292be1493d84786f68727e33ad1f91764320"
 dependencies = [
  "gix-error",
 ]
 
 [[package]]
 name = "gix-command"
-version = "0.8.0"
+version = "0.8.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b849c65a609f50d02f8a2774fe371650b3384a743c79c2a070ce0da49b7fb7da"
+checksum = "ae4bb9fa74c44c93f7238b08255f7f9afc158bafea4b95af665fa535352cd73c"
 dependencies = [
  "bstr",
  "gix-path",
@@ -2563,29 +2553,29 @@ dependencies = [
  "gix-sec",
  "memchr",
  "smallvec",
- "thiserror 2.0.18",
+ "thiserror",
  "unicode-bom",
  "winnow",
 ]
 
 [[package]]
 name = "gix-config-value"
-version = "0.17.1"
+version = "0.17.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "441a300bc3645a1f45cba495b9175f90f47256ce43f2ee161da0031e3ac77c92"
+checksum = "4378c53ec3db049919edf91ff76f56f28886a8b4b4a5a9dc633108d84afc3675"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "bstr",
  "gix-path",
  "libc",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
 name = "gix-credentials"
-version = "0.37.1"
+version = "0.37.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "38b2a34b8715e3bbd514f3d1705f5d51c4b250e5bfe506b9fb60b133c85c93d9"
+checksum = "0493f25da3ce9e8f7d925e5b64dc6d0b6109c96add422a6bcada52416448147d"
 dependencies = [
  "bstr",
  "gix-command",
@@ -2596,14 +2586,14 @@ dependencies = [
  "gix-sec",
  "gix-trace",
  "gix-url",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
 name = "gix-date"
-version = "0.15.1"
+version = "0.15.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "39acf819aa9fee65e4838a2eec5cb2506e47ebb89e02a5ab9918196e491571ea"
+checksum = "b94cdae4eb4b0f4136e3d9b3aa2d2cd03cfb5bb9b636b31263aea2df86d41543"
 dependencies = [
  "bstr",
  "gix-error",
@@ -2634,7 +2624,7 @@ dependencies = [
  "gix-worktree",
  "imara-diff 0.1.8",
  "imara-diff 0.2.0",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -2654,7 +2644,7 @@ dependencies = [
  "gix-trace",
  "gix-utils",
  "gix-worktree",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -2669,14 +2659,14 @@ dependencies = [
  "gix-path",
  "gix-ref",
  "gix-sec",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
 name = "gix-error"
-version = "0.2.1"
+version = "0.2.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2e86d01da904d4a9265def43bd42a18c5e6dc7000a73af512946ba14579c9fbd"
+checksum = "e207b971746ab724fccdfced2e4e19e854744611904a0195d3aa8fda8a110613"
 dependencies = [
  "bstr",
 ]
@@ -2698,7 +2688,7 @@ dependencies = [
  "once_cell",
  "parking_lot",
  "prodash",
- "thiserror 2.0.18",
+ "thiserror",
  "walkdir",
  "zlib-rs",
 ]
@@ -2721,7 +2711,7 @@ dependencies = [
  "gix-trace",
  "gix-utils",
  "smallvec",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -2735,7 +2725,7 @@ dependencies = [
  "gix-features",
  "gix-path",
  "gix-utils",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -2744,7 +2734,7 @@ version = "0.24.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "b03e6cd88cc0dc1eafa1fddac0fb719e4e74b6ea58dd016e71125fde4a326bee"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "bstr",
  "gix-features",
  "gix-path",
@@ -2759,7 +2749,7 @@ dependencies = [
  "faster-hex",
  "gix-features",
  "sha1-checked",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -2792,7 +2782,7 @@ version = "0.49.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "1bae54ab14e4e74d5dda60b82ea7afad7c8eb3be68283d6d5f29bd2e6d47fff7"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "bstr",
  "filetime",
  "fnv",
@@ -2811,7 +2801,7 @@ dependencies = [
  "memmap2",
  "rustix 1.1.4",
  "smallvec",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -2822,7 +2812,7 @@ checksum = "054fbd0989700c69dc5aa80bc66944f05df1e15aa7391a9e42aca7366337905f"
 dependencies = [
  "gix-tempfile",
  "gix-utils",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -2860,7 +2850,7 @@ dependencies = [
  "gix-worktree",
  "imara-diff 0.1.8",
  "nonempty",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -2869,7 +2859,7 @@ version = "0.29.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "6ea064c7595eea08fdd01c70748af747d9acc40f727b61f4c8a2145a5c5fc28c"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "gix-commitgraph",
  "gix-date",
  "gix-hash",
@@ -2894,7 +2884,7 @@ dependencies = [
  "gix-validate",
  "itoa",
  "smallvec",
- "thiserror 2.0.18",
+ "thiserror",
  "winnow",
 ]
 
@@ -2915,7 +2905,7 @@ dependencies = [
  "gix-quote",
  "parking_lot",
  "tempfile",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -2936,32 +2926,32 @@ dependencies = [
  "memmap2",
  "parking_lot",
  "smallvec",
- "thiserror 2.0.18",
+ "thiserror",
  "uluru",
 ]
 
 [[package]]
 name = "gix-packetline"
-version = "0.21.2"
+version = "0.21.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "be19313dcdb7dff75a3ce2f99be00878458295bcc3b6c7f0005591597573345c"
+checksum = "362246df440ee691699f0664cbf7006a6ece477db6734222be95e4198e5656e6"
 dependencies = [
  "bstr",
  "faster-hex",
  "gix-trace",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
 name = "gix-path"
-version = "0.11.2"
+version = "0.11.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "09c31d4373bda7fab9eb01822927b55185a378d6e1bf737e0a54c743ad806658"
+checksum = "c8fd1fe596dc393b538e1d5492c5585971a9311475b3255f7b889023df208476"
 dependencies = [
  "bstr",
  "gix-trace",
  "gix-validate",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -2970,26 +2960,26 @@ version = "0.16.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "f89611f13544ca5ebeb68a502673814ef57200df60c24a61c2ce7b96f612f08b"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "bstr",
  "gix-attributes",
  "gix-config-value",
  "gix-glob",
  "gix-path",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
 name = "gix-prompt"
-version = "0.14.1"
+version = "0.14.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0f61f6264e1f6c5a951531fe127722c7522bc02ebda80c4528286bda4642055f"
+checksum = "1de52c2570684db3eb8cebda77c1d0e3d03ddaad2d9bbb5055eba31fb9f8292a"
 dependencies = [
  "gix-command",
  "gix-config-value",
  "parking_lot",
  "rustix 1.1.4",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -3015,15 +3005,15 @@ dependencies = [
  "gix-utils",
  "maybe-async",
  "nonempty",
- "thiserror 2.0.18",
+ "thiserror",
  "winnow",
 ]
 
 [[package]]
 name = "gix-quote"
-version = "0.7.0"
+version = "0.7.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "68533db71259c8776dd4e770d2b7b98696213ecdc1f5c9e3507119e274e0c578"
+checksum = "6e97b73791a64bc0fa7dd2c5b3e551136115f97750b876ed1c952c7a7dbaf8be"
 dependencies = [
  "bstr",
  "gix-error",
@@ -3047,7 +3037,7 @@ dependencies = [
  "gix-utils",
  "gix-validate",
  "memmap2",
- "thiserror 2.0.18",
+ "thiserror",
  "winnow",
 ]
 
@@ -3064,7 +3054,7 @@ dependencies = [
  "gix-revision",
  "gix-validate",
  "smallvec",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -3073,7 +3063,7 @@ version = "0.43.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7c08f1ec5d1e6a524f8ba291c41f0ccaef64e48ed0e8cf790b3461cae45f6d3d"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "bstr",
  "gix-commitgraph",
  "gix-date",
@@ -3099,16 +3089,16 @@ dependencies = [
  "gix-hashtable",
  "gix-object",
  "smallvec",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
 name = "gix-sec"
-version = "0.13.2"
+version = "0.13.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bf82ae037de9c62850ce67beaa92ec8e3e17785ea307cdde7618edc215603b4f"
+checksum = "283f4a746c9bde8550be63e6f961ff4651f412ca12666e8f5615f39464960ab9"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "gix-path",
  "libc",
  "windows-sys 0.61.2",
@@ -3124,7 +3114,7 @@ dependencies = [
  "gix-hash",
  "gix-lock",
  "nonempty",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -3147,7 +3137,7 @@ dependencies = [
  "gix-pathspec",
  "gix-worktree",
  "portable-atomic",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -3162,7 +3152,7 @@ dependencies = [
  "gix-pathspec",
  "gix-refspec",
  "gix-url",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -3171,7 +3161,7 @@ version = "21.0.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "d22227f6b203f511ff451c33c89899e87e4f571fc596b06f68e6e613a6508528"
 dependencies = [
- "dashmap 6.1.0",
+ "dashmap",
  "gix-fs",
  "libc",
  "parking_lot",
@@ -3182,9 +3172,9 @@ dependencies = [
 
 [[package]]
 name = "gix-trace"
-version = "0.1.18"
+version = "0.1.19"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f69a13643b8437d4ca6845e08143e847a36ca82903eed13303475d0ae8b162e0"
+checksum = "6f23569e55f2ffaf958617353b9734a7d52a7c19c439eeaa5e3efc217fd2270e"
 
 [[package]]
 name = "gix-transport"
@@ -3202,7 +3192,7 @@ dependencies = [
  "gix-quote",
  "gix-sec",
  "gix-url",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -3211,7 +3201,7 @@ version = "0.55.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "963dc2afcdb611092aa587c3f9365e749ac0a0892ff27662dbc75f26c953fbec"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "gix-commitgraph",
  "gix-date",
  "gix-hash",
@@ -3219,26 +3209,26 @@ dependencies = [
  "gix-object",
  "gix-revwalk",
  "smallvec",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
 name = "gix-url"
-version = "0.35.2"
+version = "0.35.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d28e8af3d42581190da884f013caf254d2fd4d6ab102408f08d21bfa11de6c8d"
+checksum = "1a61ead12e33fa52ae92b207ee27554f646a8e7a3dad8b78da1582ec91eda0a6"
 dependencies = [
  "bstr",
  "gix-path",
  "percent-encoding",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
 name = "gix-utils"
-version = "0.3.1"
+version = "0.3.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "befcdbdfb1238d2854591f760a48711bed85e72d80a10e8f2f93f656746ef7c5"
+checksum = "4e477b4f07a6e8da4ba791c53c858102959703c60d70f199932010d5b94adb2c"
 dependencies = [
  "bstr",
  "fastrand",
@@ -3247,9 +3237,9 @@ dependencies = [
 
 [[package]]
 name = "gix-validate"
-version = "0.11.0"
+version = "0.11.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0ec1eff98d91941f47766367cba1be746bab662bad761d9891ae6f7882f7840b"
+checksum = "e26ac2602b43eadfdca0560b81d3341944162a3c9f64ccdeef8fc501ad80dad5"
 dependencies = [
  "bstr",
 ]
@@ -3287,7 +3277,7 @@ dependencies = [
  "gix-path",
  "gix-worktree",
  "io-close",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -3316,9 +3306,9 @@ checksum = "0cc23270f6e1808e30a928bdc84dea0b9b4136a8bc82338574f23baf47bbd280"
 
 [[package]]
 name = "h2"
-version = "0.4.13"
+version = "0.4.14"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2f44da3a8150a6703ed5d34e164b875fd14c2cdab9af1252a9a1020bde2bdc54"
+checksum = "171fefbc92fe4a4de27e0698d6a5b392d6a0e333506bc49133760b3bcf948733"
 dependencies = [
  "atomic-waker",
  "bytes",
@@ -3326,7 +3316,7 @@ dependencies = [
  "futures-core",
  "futures-sink",
  "http",
- "indexmap 2.13.0",
+ "indexmap 2.14.0",
  "slab",
  "tokio",
  "tokio-util",
@@ -3388,6 +3378,12 @@ dependencies = [
  "foldhash 0.2.0",
 ]
 
+[[package]]
+name = "hashbrown"
+version = "0.17.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "4f467dd6dccf739c208452f8014c75c18bb8301b050ad1cfb27153803edb0f51"
+
 [[package]]
 name = "heapless"
 version = "0.8.0"
@@ -3427,10 +3423,10 @@ dependencies = [
  "indicatif 0.17.11",
  "libc",
  "log",
- "rand 0.9.2",
+ "rand 0.9.4",
  "serde",
  "serde_json",
- "thiserror 2.0.18",
+ "thiserror",
  "ureq",
  "windows-sys 0.60.2",
 ]
@@ -3509,9 +3505,9 @@ checksum = "135b12329e5e3ce057a9f972339ea52bc954fe1e9358ef27f95e89716fbc5424"
 
 [[package]]
 name = "hyper"
-version = "1.8.1"
+version = "1.9.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2ab2d4f250c3d7b1c9fcdff1cece94ea4e2dfbec68614f7b87cb205f24ca9d11"
+checksum = "6299f016b246a94207e63da54dbe807655bf9e00044f73ded42c3ac5305fbcca"
 dependencies = [
  "atomic-waker",
  "bytes",
@@ -3524,7 +3520,6 @@ dependencies = [
  "httpdate",
  "itoa",
  "pin-project-lite",
- "pin-utils",
  "smallvec",
  "tokio",
  "want",
@@ -3532,16 +3527,15 @@ dependencies = [
 
 [[package]]
 name = "hyper-rustls"
-version = "0.27.7"
+version = "0.27.9"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e3c93eb611681b207e1fe55d5a71ecf91572ec8a6705cdb6857f7d8d5242cf58"
+checksum = "33ca68d021ef39cf6463ab54c1d0f5daf03377b70561305bb89a8f83aab66e0f"
 dependencies = [
  "http",
  "hyper",
  "hyper-util",
  "rustls",
  "rustls-native-certs",
- "rustls-pki-types",
  "tokio",
  "tokio-rustls",
  "tower-service",
@@ -3607,12 +3601,13 @@ dependencies = [
 
 [[package]]
 name = "icu_collections"
-version = "2.1.1"
+version = "2.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4c6b649701667bbe825c3b7e6388cb521c23d88644678e83c0c4d0a621a34b43"
+checksum = "2984d1cd16c883d7935b9e07e44071dca8d917fd52ecc02c04d5fa0b5a3f191c"
 dependencies = [
  "displaydoc",
  "potential_utf",
+ "utf8_iter",
  "yoke",
  "zerofrom",
  "zerovec",
@@ -3620,9 +3615,9 @@ dependencies = [
 
 [[package]]
 name = "icu_locale_core"
-version = "2.1.1"
+version = "2.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "edba7861004dd3714265b4db54a3c390e880ab658fec5f7db895fae2046b5bb6"
+checksum = "92219b62b3e2b4d88ac5119f8904c10f8f61bf7e95b640d25ba3075e6cac2c29"
 dependencies = [
  "displaydoc",
  "litemap",
@@ -3633,9 +3628,9 @@ dependencies = [
 
 [[package]]
 name = "icu_normalizer"
-version = "2.1.1"
+version = "2.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5f6c8828b67bf8908d82127b2054ea1b4427ff0230ee9141c54251934ab1b599"
+checksum = "c56e5ee99d6e3d33bd91c5d85458b6005a22140021cc324cea84dd0e72cff3b4"
 dependencies = [
  "icu_collections",
  "icu_normalizer_data",
@@ -3647,15 +3642,15 @@ dependencies = [
 
 [[package]]
 name = "icu_normalizer_data"
-version = "2.1.1"
+version = "2.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7aedcccd01fc5fe81e6b489c15b247b8b0690feb23304303a9e560f37efc560a"
+checksum = "da3be0ae77ea334f4da67c12f149704f19f81d1adf7c51cf482943e84a2bad38"
 
 [[package]]
 name = "icu_properties"
-version = "2.1.2"
+version = "2.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "020bfc02fe870ec3a66d93e677ccca0562506e5872c650f893269e08615d74ec"
+checksum = "bee3b67d0ea5c2cca5003417989af8996f8604e34fb9ddf96208a033901e70de"
 dependencies = [
  "icu_collections",
  "icu_locale_core",
@@ -3667,15 +3662,15 @@ dependencies = [
 
 [[package]]
 name = "icu_properties_data"
-version = "2.1.2"
+version = "2.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "616c294cf8d725c6afcd8f55abc17c56464ef6211f9ed59cccffe534129c77af"
+checksum = "8e2bbb201e0c04f7b4b3e14382af113e17ba4f63e2c9d2ee626b720cbce54a14"
 
 [[package]]
 name = "icu_provider"
-version = "2.1.1"
+version = "2.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "85962cf0ce02e1e0a629cc34e7ca3e373ce20dda4c4d7294bbd0bf1fdb59e614"
+checksum = "139c4cf31c8b5f33d7e199446eff9c1e02decfc2f0eec2c8d71f65befa45b421"
 dependencies = [
  "displaydoc",
  "icu_locale_core",
@@ -3711,9 +3706,9 @@ dependencies = [
 
 [[package]]
 name = "idna_adapter"
-version = "1.2.1"
+version = "1.2.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3acae9609540aa318d1bc588455225fb2085b9ed0c4f6bd0d9d5bcd86f1a0344"
+checksum = "cb68373c0d6620ef8105e855e7745e18b0d00d3bdb07fb532e434244cdb9a714"
 dependencies = [
  "icu_normalizer",
  "icu_properties",
@@ -3751,12 +3746,12 @@ dependencies = [
 
 [[package]]
 name = "indexmap"
-version = "2.13.0"
+version = "2.14.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7714e70437a7dc3ac8eb7e6f8df75fd8eb422675fc7678aff7364301092b1017"
+checksum = "d466e9454f08e4a911e14806c24e16fba1b4c121d1ea474396f396069cf949d9"
 dependencies = [
  "equivalent",
- "hashbrown 0.16.1",
+ "hashbrown 0.17.0",
  "serde",
  "serde_core",
 ]
@@ -3805,9 +3800,9 @@ checksum = "d98f6fed1fde3f8c21bc40a1abb88dd75e67924f9cffc3ef95607bad8017f8e2"
 
 [[package]]
 name = "iri-string"
-version = "0.7.11"
+version = "0.7.12"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d8e7418f59cc01c88316161279a7f665217ae316b388e58a0d10e29f54f1e5eb"
+checksum = "25e659a4bb38e810ebc252e53b5814ff908a8c58c2a9ce2fae1bbec24cbf4e20"
 dependencies = [
  "memchr",
  "serde",
@@ -3821,7 +3816,7 @@ checksum = "3640c1c38b8e4e43584d8df18be5fc6b0aa314ce6ebf51b53313d4306cca8e46"
 dependencies = [
  "hermit-abi",
  "libc",
- "windows-sys 0.52.0",
+ "windows-sys 0.61.2",
 ]
 
 [[package]]
@@ -3862,9 +3857,9 @@ checksum = "8f42a60cbdf9a97f5d2305f08a87dc4e09308d1276d28c869c684d7777685682"
 
 [[package]]
 name = "jiff"
-version = "0.2.23"
+version = "0.2.24"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1a3546dc96b6d42c5f24902af9e2538e82e39ad350b0c766eb3fbf2d8f3d8359"
+checksum = "f00b5dbd620d61dfdcb6007c9c1f6054ebd75319f163d886a9055cec1155073d"
 dependencies = [
  "jiff-static",
  "jiff-tzdb-platform",
@@ -3872,14 +3867,14 @@ dependencies = [
  "portable-atomic",
  "portable-atomic-util",
  "serde_core",
- "windows-sys 0.52.0",
+ "windows-sys 0.61.2",
 ]
 
 [[package]]
 name = "jiff-static"
-version = "0.2.23"
+version = "0.2.24"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2a8c8b344124222efd714b73bb41f8b5120b27a7cc1c75593a6ff768d9d05aa4"
+checksum = "e000de030ff8022ea1da3f466fbb0f3a809f5e51ed31f6dd931c35181ad8e6d7"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -3903,27 +3898,32 @@ dependencies = [
 
 [[package]]
 name = "jni"
-version = "0.21.1"
+version = "0.22.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1a87aa2bb7d2af34197c04845522473242e1aa17c12f4935d5856491a7fb8c97"
+checksum = "5efd9a482cf3a427f00d6b35f14332adc7902ce91efb778580e180ff90fa3498"
 dependencies = [
- "cesu8",
  "cfg-if",
  "combine",
- "jni-sys 0.3.1",
+ "jni-macros",
+ "jni-sys",
  "log",
- "thiserror 1.0.69",
+ "simd_cesu8",
+ "thiserror",
  "walkdir",
- "windows-sys 0.45.0",
+ "windows-link",
 ]
 
 [[package]]
-name = "jni-sys"
-version = "0.3.1"
+name = "jni-macros"
+version = "0.22.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "41a652e1f9b6e0275df1f15b32661cf0d4b78d4d87ddec5e0c3c20f097433258"
+checksum = "a00109accc170f0bdb141fed3e393c565b6f5e072365c3bd58f5b062591560a3"
 dependencies = [
- "jni-sys 0.4.1",
+ "proc-macro2",
+ "quote",
+ "rustc_version",
+ "simd_cesu8",
+ "syn 2.0.117",
 ]
 
 [[package]]
@@ -3957,19 +3957,21 @@ dependencies = [
 
 [[package]]
 name = "js-sys"
-version = "0.3.91"
+version = "0.3.97"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b49715b7073f385ba4bc528e5747d02e66cb39c6146efb66b781f131f0fb399c"
+checksum = "a1840c94c045fbcf8ba2812c95db44499f7c64910a912551aaaa541decebcacf"
 dependencies = [
+ "cfg-if",
+ "futures-util",
  "once_cell",
  "wasm-bindgen",
 ]
 
 [[package]]
 name = "jsonb"
-version = "0.5.5"
+version = "0.5.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2a901f06163d352fbe41c3c2ff5e08b75330a003cc941e988fb501022f5421e6"
+checksum = "eb98fb29636087c40ad0d1274d9a30c0c1e83e03ae93f6e7e89247b37fcc6953"
 dependencies = [
  "byteorder",
  "ethnum",
@@ -3979,10 +3981,10 @@ dependencies = [
  "nom 8.0.0",
  "num-traits",
  "ordered-float",
- "rand 0.9.2",
- "ryu",
+ "rand 0.9.4",
  "serde",
  "serde_json",
+ "zmij",
 ]
 
 [[package]]
@@ -3996,9 +3998,9 @@ dependencies = [
 
 [[package]]
 name = "lance"
-version = "3.0.1"
+version = "4.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "95c5ce428fda0721f5c48bfde17a1921c4da2d2142b2f46a16c89abf5fce8003"
+checksum = "efe6c3ddd79cdfd2b7e1c23cafae52806906bc40fbd97de9e8cf2f8c7a75fc04"
 dependencies = [
  "arrow",
  "arrow-arith",
@@ -4016,7 +4018,7 @@ dependencies = [
  "bytes",
  "chrono",
  "crossbeam-skiplist",
- "dashmap 6.1.0",
+ "dashmap",
  "datafusion",
  "datafusion-expr",
  "datafusion-functions",
@@ -4045,7 +4047,7 @@ dependencies = [
  "pin-project",
  "prost",
  "prost-types",
- "rand 0.9.2",
+ "rand 0.9.4",
  "roaring",
  "semver",
  "serde",
@@ -4062,9 +4064,9 @@ dependencies = [
 
 [[package]]
 name = "lance-arrow"
-version = "3.0.1"
+version = "4.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c9fdaf99863fa0d631e422881e88be4837d8b82f36a87143d723a9d285acec4b"
+checksum = "5d9f5d95bdda2a2b790f1fb8028b5b6dcf661abeb3133a8bca0f3d24b054af87"
 dependencies = [
  "arrow-array",
  "arrow-buffer",
@@ -4079,14 +4081,14 @@ dependencies = [
  "half",
  "jsonb",
  "num-traits",
- "rand 0.9.2",
+ "rand 0.9.4",
 ]
 
 [[package]]
 name = "lance-bitpacking"
-version = "3.0.1"
+version = "4.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "866b1634d38d94e8ab86fbcf238ac82dc8a5f72a4a6a90525f29899772e7cc7f"
+checksum = "f827d6ab9f8f337a9509d5ad66a12f3314db8713868260521c344ef6135eb4e4"
 dependencies = [
  "arrayref",
  "paste",
@@ -4095,9 +4097,9 @@ dependencies = [
 
 [[package]]
 name = "lance-core"
-version = "3.0.1"
+version = "4.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "977c29f4e48c201c2806fe6ae117b65d0287eda236acd07357b556a54b0d5c5a"
+checksum = "0f1e25df6a79bf72ee6bcde0851f19b1cd36c5848c1b7db83340882d3c9fdecb"
 dependencies = [
  "arrow-array",
  "arrow-buffer",
@@ -4120,7 +4122,7 @@ dependencies = [
  "object_store",
  "pin-project",
  "prost",
- "rand 0.9.2",
+ "rand 0.9.4",
  "roaring",
  "serde_json",
  "snafu 0.9.0",
@@ -4134,9 +4136,9 @@ dependencies = [
 
 [[package]]
 name = "lance-datafusion"
-version = "3.0.1"
+version = "4.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0ccc72695473f4207df4c6df3b347a63e84c32c0bc36bf42a7d86e8a7c0c67e2"
+checksum = "93146de8ae720cb90edef81c2f2d0a1b065fc2f23ecff2419546f389b0fa70a4"
 dependencies = [
  "arrow",
  "arrow-array",
@@ -4166,9 +4168,9 @@ dependencies = [
 
 [[package]]
 name = "lance-datagen"
-version = "3.0.1"
+version = "4.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8fe84d76944acd834ded14d7562663af995556e0c6594f4b4ac69b0183f99c1a"
+checksum = "ccec8ce4d8e0a87a99c431dab2364398029f2ffb649c1a693c60c79e05ed30dd"
 dependencies = [
  "arrow",
  "arrow-array",
@@ -4178,7 +4180,7 @@ dependencies = [
  "futures",
  "half",
  "hex",
- "rand 0.9.2",
+ "rand 0.9.4",
  "rand_distr 0.5.1",
  "rand_xoshiro",
  "random_word",
@@ -4186,9 +4188,9 @@ dependencies = [
 
 [[package]]
 name = "lance-encoding"
-version = "3.0.1"
+version = "4.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "be1007242188e5d53c98717e7f2cb340dc80eb9c94c2b935587598919b3a36bd"
+checksum = "5c1aec0bbbac6bce829bc10f1ba066258126100596c375fb71908ecf11c2c2a5"
 dependencies = [
  "arrow-arith",
  "arrow-array",
@@ -4214,7 +4216,7 @@ dependencies = [
  "prost",
  "prost-build",
  "prost-types",
- "rand 0.9.2",
+ "rand 0.9.4",
  "snafu 0.9.0",
  "strum",
  "tokio",
@@ -4225,9 +4227,9 @@ dependencies = [
 
 [[package]]
 name = "lance-file"
-version = "3.0.1"
+version = "4.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f80088e418941f39cf5599d166ae1a6ef498cc2d967652a0692477d4871a9277"
+checksum = "14a8c548804f5b17486dc2d3282356ed1957095a852780283bc401fdd69e9075"
 dependencies = [
  "arrow-arith",
  "arrow-array",
@@ -4259,9 +4261,9 @@ dependencies = [
 
 [[package]]
 name = "lance-index"
-version = "3.0.1"
+version = "4.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e0011daf1ddde99becffd2ae235ad324576736a526c54ffbc4d7e583872f1215"
+checksum = "2da212f0090ea59f79ac3686660f596520c167fe1cb5f408900cf71d215f0e03"
 dependencies = [
  "arrow",
  "arrow-arith",
@@ -4275,6 +4277,7 @@ dependencies = [
  "bitpacking",
  "bitvec",
  "bytes",
+ "chrono",
  "crossbeam-queue",
  "datafusion",
  "datafusion-common",
@@ -4305,7 +4308,7 @@ dependencies = [
  "prost",
  "prost-build",
  "prost-types",
- "rand 0.9.2",
+ "rand 0.9.4",
  "rand_distr 0.5.1",
  "rangemap",
  "rayon",
@@ -4324,9 +4327,9 @@ dependencies = [
 
 [[package]]
 name = "lance-io"
-version = "3.0.1"
+version = "4.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "cfa8a74e93753d19a27ce3adaeb99e31227df13ad5926dd43572be76b43dd284"
+checksum = "41d958eb4b56f03bbe0f5f85eb2b4e9657882812297b6f711f201ffc995f259f"
 dependencies = [
  "arrow",
  "arrow-arith",
@@ -4352,7 +4355,7 @@ dependencies = [
  "path_abs",
  "pin-project",
  "prost",
- "rand 0.9.2",
+ "rand 0.9.4",
  "serde",
  "snafu 0.9.0",
  "tempfile",
@@ -4363,9 +4366,9 @@ dependencies = [
 
 [[package]]
 name = "lance-linalg"
-version = "3.0.1"
+version = "4.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6e2d8da8f6b8dd37ab3b8199896ee265817f86232e3727c0b0eeb3c9093b64d9"
+checksum = "0285b70da35def7ed95e150fae1d5308089554e1290470403ed3c50cb235bc5e"
 dependencies = [
  "arrow-array",
  "arrow-buffer",
@@ -4376,28 +4379,29 @@ dependencies = [
  "lance-arrow",
  "lance-core",
  "num-traits",
- "rand 0.9.2",
+ "rand 0.9.4",
 ]
 
 [[package]]
 name = "lance-namespace"
-version = "3.0.1"
+version = "4.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f176e427d9c35938d8a7097876114bc35dfd280b06077779753f2effe3e86aab"
+checksum = "5f78e2a828b654e062a495462c6e3eb4fcf0e7e907d761b8f217fc09ccd3ceac"
 dependencies = [
  "arrow",
  "async-trait",
  "bytes",
  "lance-core",
  "lance-namespace-reqwest-client",
+ "serde",
  "snafu 0.9.0",
 ]
 
 [[package]]
 name = "lance-namespace-impls"
-version = "3.0.1"
+version = "4.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "663c32086ecfab311acb0813c65a4bb352a5b648ccf8b513c24697ce8d412039"
+checksum = "a2392314f3da38f00d166295e44244208a65ccfc256e274fa8631849fc3f4d94"
 dependencies = [
  "arrow",
  "arrow-ipc",
@@ -4414,7 +4418,7 @@ dependencies = [
  "lance-table",
  "log",
  "object_store",
- "rand 0.9.2",
+ "rand 0.9.4",
  "serde_json",
  "snafu 0.9.0",
  "tokio",
@@ -4423,9 +4427,9 @@ dependencies = [
 
 [[package]]
 name = "lance-namespace-reqwest-client"
-version = "0.5.4"
+version = "0.6.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "df9008f9825066088178c10599130c8bb0b9c79a39a479e8c51201620c43864a"
+checksum = "ee2e48de899e2931afb67fcddd0a08e439bf5d8b6ea2a2ed9cb8f4df669bd5cc"
 dependencies = [
  "reqwest 0.12.28",
  "serde",
@@ -4436,9 +4440,9 @@ dependencies = [
 
 [[package]]
 name = "lance-table"
-version = "3.0.1"
+version = "4.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "aa189b3081481a97b64cf1161297947a63b8adb941b1950989d0269858703a43"
+checksum = "3df9c4adca3eb2074b3850432a9fb34248a3d90c3d6427d158b13ff9355664ee"
 dependencies = [
  "arrow",
  "arrow-array",
@@ -4460,7 +4464,7 @@ dependencies = [
  "prost",
  "prost-build",
  "prost-types",
- "rand 0.9.2",
+ "rand 0.9.4",
  "rangemap",
  "roaring",
  "semver",
@@ -4475,22 +4479,22 @@ dependencies = [
 
 [[package]]
 name = "lance-testing"
-version = "3.0.1"
+version = "4.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "79a6f4ab0788ee82893bac5de4ff0d0d88bba96de87db4b6e18b1883616d4dbe"
+checksum = "7ed7119bdd6983718387b4ac44af873a165262ca94f181b104cd6f97912eb3bf"
 dependencies = [
  "arrow-array",
  "arrow-schema",
  "lance-arrow",
  "num-traits",
- "rand 0.9.2",
+ "rand 0.9.4",
 ]
 
 [[package]]
 name = "lancedb"
-version = "0.27.1"
+version = "0.27.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b79dd30fddeb0f21d090c502f937aa4862b5a352e74c47e35c6c76fec9f31534"
+checksum = "ce0f4d7f739dc30608fe8b202cbb40986c2937e1a5a189f98fb06d7b8543156a"
 dependencies = [
  "ahash",
  "arrow",
@@ -4535,7 +4539,7 @@ dependencies = [
  "num-traits",
  "object_store",
  "pin-project",
- "rand 0.9.2",
+ "rand 0.9.4",
  "regex",
  "semver",
  "serde",
@@ -4625,9 +4629,9 @@ dependencies = [
 
 [[package]]
 name = "libc"
-version = "0.2.183"
+version = "0.2.186"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b5b646652bf6661599e1da8901b3b9522896f01e736bad5f723fe7a3a27f899d"
+checksum = "68ab91017fe16c622486840e4c83c9a37afeff978bd239b5293d61ece587de66"
 
 [[package]]
 name = "libm"
@@ -4637,21 +4641,21 @@ checksum = "b6d2cec3eae94f9f509c767b45932f1ada8350c4bdb85af2fcab4a3c14807981"
 
 [[package]]
 name = "libredox"
-version = "0.1.14"
+version = "0.1.16"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1744e39d1d6a9948f4f388969627434e31128196de472883b39f148769bfe30a"
+checksum = "e02f3bb43d335493c96bf3fd3a321600bf6bd07ed34bc64118e9293bdffea46c"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "libc",
  "plain",
- "redox_syscall 0.7.3",
+ "redox_syscall 0.7.5",
 ]
 
 [[package]]
 name = "libz-sys"
-version = "1.1.25"
+version = "1.1.28"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d52f4c29e2a68ac30c9087e1b772dc9f44a2b66ed44edf2266cf2be9b03dafc1"
+checksum = "fc3a226e576f50782b3305c5ccf458698f92798987f551c6a02efe8276721e22"
 dependencies = [
  "cc",
  "libc",
@@ -4673,9 +4677,9 @@ checksum = "32a66949e030da00e8c7d4434b251670a91556f4144941d37452769c25d58a53"
 
 [[package]]
 name = "litemap"
-version = "0.8.1"
+version = "0.8.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6373607a59f0be73a39b6fe456b8192fcc3585f602af20751600e974dd455e77"
+checksum = "92daf443525c4cce67b150400bc2316076100ce0b3686209eb8cf3c31612e6f0"
 
 [[package]]
 name = "lock_api"
@@ -4721,16 +4725,16 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "112b39cec0b298b6c1999fee3e31427f74f676e4cb9879ed1a121b43661a4154"
 
 [[package]]
-name = "lsp-types"
-version = "0.94.1"
+name = "ls-types"
+version = "0.0.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c66bfd44a06ae10647fe3f8214762e9369fd4248df1350924b4ef9e770a85ea1"
+checksum = "896e16b8e17d8732b9efe4d5b66cb0cc162b3023a2d8122f2aea6f7f185e0a67"
 dependencies = [
- "bitflags 1.3.2",
+ "bitflags",
+ "fluent-uri",
+ "percent-encoding",
  "serde",
  "serde_json",
- "serde_repr",
- "url",
 ]
 
 [[package]]
@@ -4890,9 +4894,9 @@ dependencies = [
 
 [[package]]
 name = "mio"
-version = "1.1.1"
+version = "1.2.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a69bcab0ad47271a0234d9422b131806bf3968021e5dc9328caf2d4cd58557fc"
+checksum = "50b7e5b27aa02a74bac8c3f23f448f8d87ff11f92d3aac1a6ed369ee08cc56c1"
 dependencies = [
  "libc",
  "wasi",
@@ -5033,7 +5037,7 @@ version = "0.30.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "74523f3a35e05aba87a1d978330aef40f67b0304ac79c1c00b294c9830543db6"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "cfg-if",
  "cfg_aliases",
  "libc",
@@ -5070,7 +5074,7 @@ version = "0.50.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7957b9740744892f114936ab4a57b3f487491bbeafaf8083688b16841a4240e5"
 dependencies = [
- "windows-sys 0.60.2",
+ "windows-sys 0.61.2",
 ]
 
 [[package]]
@@ -5094,9 +5098,9 @@ dependencies = [
 
 [[package]]
 name = "num-conv"
-version = "0.2.0"
+version = "0.2.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "cf97ec579c3c42f953ef76dbf8d55ac91fb219dde70e49aa4a6b7d74e9919050"
+checksum = "c6673768db2d862beb9b39a78fdcb1a69439615d5794a1be50caa9bc92c81967"
 
 [[package]]
 name = "num-integer"
@@ -5148,7 +5152,7 @@ dependencies = [
  "itertools 0.14.0",
  "parking_lot",
  "percent-encoding",
- "thiserror 2.0.18",
+ "thiserror",
  "tokio",
  "tracing",
  "url",
@@ -5183,11 +5187,11 @@ checksum = "cfe21416a02c693fb9f980befcb230ecc70b0b3d1cc4abf88b9675c4c1457f0c"
 
 [[package]]
 name = "onig"
-version = "6.5.1"
+version = "6.5.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "336b9c63443aceef14bea841b899035ae3abe89b7c486aaf4c5bd8aafedac3f0"
+checksum = "0cc3cbf698f9438986c11a880c90a6d04b9de27575afd28bbf45b154b6c709e2"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "libc",
  "once_cell",
  "onig_sys",
@@ -5195,9 +5199,9 @@ dependencies = [
 
 [[package]]
 name = "onig_sys"
-version = "69.9.1"
+version = "69.9.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c7f86c6eef3d6df15f23bcfb6af487cbd2fed4e5581d58d5bf1f5f8b7f6727dc"
+checksum = "1e68317604e77e53b85896388e1a803c1d21b74c899ec9e5e1112db90735edd7"
 dependencies = [
  "cc",
  "pkg-config",
@@ -5217,9 +5221,9 @@ checksum = "7c87def4c32ab89d880effc9e097653c8da5d6ef28e6b539d313baaacfbafcbe"
 
 [[package]]
 name = "openssl-sys"
-version = "0.9.112"
+version = "0.9.115"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "57d55af3b3e226502be1526dfdba67ab0e9c96fc293004e79576b2b9edb0dbdb"
+checksum = "158fe5b292746440aa6e7a7e690e55aeb72d41505e2804c23c6973ad0e9c9781"
 dependencies = [
  "cc",
  "libc",
@@ -5235,9 +5239,9 @@ checksum = "04744f49eae99ab78e0d5c0b603ab218f515ea8cfe5a456d7629ad883a3b6e7d"
 
 [[package]]
 name = "ordered-float"
-version = "5.1.0"
+version = "5.3.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7f4779c6901a562440c3786d08192c6fbda7c1c2060edd10006b05ee35d10f2d"
+checksum = "b7d950ca161dc355eaf28f82b11345ed76c6e1f6eb1f4f4479e0323b9e2fbd0e"
 dependencies = [
  "num-traits",
 ]
@@ -5328,7 +5332,7 @@ checksum = "8701b58ea97060d5e5b155d383a69952a60943f0e6dfe30b04c287beb0b27455"
 dependencies = [
  "fixedbitset",
  "hashbrown 0.15.5",
- "indexmap 2.13.0",
+ "indexmap 2.14.0",
  "serde",
 ]
 
@@ -5352,18 +5356,18 @@ dependencies = [
 
 [[package]]
 name = "pin-project"
-version = "1.1.11"
+version = "1.1.12"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f1749c7ed4bcaf4c3d0a3efc28538844fb29bcdd7d2b67b2be7e20ba861ff517"
+checksum = "cbf0d9e68100b3a7989b4901972f265cd542e560a3a8a724e1e20322f4d06ce9"
 dependencies = [
  "pin-project-internal",
 ]
 
 [[package]]
 name = "pin-project-internal"
-version = "1.1.11"
+version = "1.1.12"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d9b20ed30f105399776b9c883e68e536ef602a16ae6f596d2c473591d6ad64c6"
+checksum = "a990e22f43e84855daf260dded30524ef4a9021cc7541c26540500a50b624389"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -5376,17 +5380,11 @@ version = "0.2.17"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "a89322df9ebe1c1578d689c92318e070967d1042b512afbe49518723f4e6d5cd"
 
-[[package]]
-name = "pin-utils"
-version = "0.1.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8b870d8c151b6f2fb93e84a13146138f05d02ed11c7e7c54f8826aaaf7c9f184"
-
 [[package]]
 name = "pkg-config"
-version = "0.3.32"
+version = "0.3.33"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "7edddbd0b52d732b21ad9a5fab5c704c14cd949e5e9a1ec5929a24fded1b904c"
+checksum = "19f132c84eca552bf34cab8ec81f1c1dcc229b811638f9d283dceabe58c5569e"
 
 [[package]]
 name = "plain"
@@ -5402,18 +5400,18 @@ checksum = "c33a9471896f1c69cecef8d20cbe2f7accd12527ce60845ff44c153bb2a21b49"
 
 [[package]]
 name = "portable-atomic-util"
-version = "0.2.6"
+version = "0.2.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "091397be61a01d4be58e7841595bd4bfedb15f1cd54977d79b8271e94ed799a3"
+checksum = "c2a106d1259c23fac8e543272398ae0e3c0b8d33c88ed73d0cc71b0f1d902618"
 dependencies = [
  "portable-atomic",
 ]
 
 [[package]]
 name = "potential_utf"
-version = "0.1.4"
+version = "0.1.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b73949432f5e2a09657003c25bca5e19a0e9c84f8058ca374f49e0ebe605af77"
+checksum = "0103b1cef7ec0cf76490e969665504990193874ea05c85ff9bab8b911d0a0564"
 dependencies = [
  "zerovec",
 ]
@@ -5528,7 +5526,7 @@ dependencies = [
  "rustc-hash",
  "rustls",
  "socket2",
- "thiserror 2.0.18",
+ "thiserror",
  "tokio",
  "tracing",
  "web-time",
@@ -5544,13 +5542,13 @@ dependencies = [
  "bytes",
  "getrandom 0.3.4",
  "lru-slab",
- "rand 0.9.2",
+ "rand 0.9.4",
  "ring",
  "rustc-hash",
  "rustls",
  "rustls-pki-types",
  "slab",
- "thiserror 2.0.18",
+ "thiserror",
  "tinyvec",
  "tracing",
  "web-time",
@@ -5609,9 +5607,9 @@ dependencies = [
 
 [[package]]
 name = "rand"
-version = "0.8.5"
+version = "0.8.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "34af8d1a0e25924bc5b7c43c079c942339d8f0a8b57c39049bef581b46327404"
+checksum = "5ca0ecfa931c29007047d1bc58e623ab12e5590e8c7cc53200d5202b69266d8a"
 dependencies = [
  "libc",
  "rand_chacha 0.3.1",
@@ -5620,9 +5618,9 @@ dependencies = [
 
 [[package]]
 name = "rand"
-version = "0.9.2"
+version = "0.9.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6db2770f06117d490610c7488547d543617b21bfa07796d7a12f6f1bd53850d1"
+checksum = "44c5af06bb1b7d3216d91932aed5265164bf384dc89cd6ba05cf59a35f5f76ea"
 dependencies = [
  "rand_chacha 0.9.0",
  "rand_core 0.9.5",
@@ -5673,7 +5671,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "32cb0b9bc82b0a0876c2dd994a7e7a2683d3e7390ca40e6886785ef0c7e3ee31"
 dependencies = [
  "num-traits",
- "rand 0.8.5",
+ "rand 0.8.6",
 ]
 
 [[package]]
@@ -5683,7 +5681,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "6a8615d50dcf34fa31f7ab52692afec947c4dd0ab803cc87cb3b0b4570ff7463"
 dependencies = [
  "num-traits",
- "rand 0.9.2",
+ "rand 0.9.4",
 ]
 
 [[package]]
@@ -5704,7 +5702,7 @@ dependencies = [
  "ahash",
  "brotli",
  "paste",
- "rand 0.9.2",
+ "rand 0.9.4",
  "unicase",
 ]
 
@@ -5722,9 +5720,9 @@ checksum = "60a357793950651c4ed0f3f52338f53b2f809f32d83a07f72909fa13e4c6c1e3"
 
 [[package]]
 name = "rayon"
-version = "1.11.0"
+version = "1.12.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "368f01d005bf8fd9b1206fb6fa653e6c4a81ceb1466406b81792d87c5677a58f"
+checksum = "fb39b166781f92d482534ef4b4b1b2568f42613b53e5b6c160e24cfbfa30926d"
 dependencies = [
  "either",
  "rayon-core",
@@ -5757,16 +5755,16 @@ version = "0.5.18"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "ed2bf2547551a7053d6fdfafda3f938979645c44812fbfcda098faae3f1a362d"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
 ]
 
 [[package]]
 name = "redox_syscall"
-version = "0.7.3"
+version = "0.7.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6ce70a74e890531977d37e532c34d45e9055d2409ed08ddba14529471ed0be16"
+checksum = "4666a1a60d8412eab19d94f6d13dcc9cea0a5ef4fdf6a5db306537413c661b1b"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
 ]
 
 [[package]]
@@ -5777,7 +5775,7 @@ checksum = "a4e608c6638b9c18977b00b475ac1f28d14e84b27d8d42f70e0bf1e3dec127ac"
 dependencies = [
  "getrandom 0.2.17",
  "libredox",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -5864,7 +5862,7 @@ dependencies = [
  "tokio",
  "tokio-rustls",
  "tokio-util",
- "tower 0.5.3",
+ "tower",
  "tower-http",
  "tower-service",
  "url",
@@ -5876,9 +5874,9 @@ dependencies = [
 
 [[package]]
 name = "reqwest"
-version = "0.13.2"
+version = "0.13.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ab3f43e3283ab1488b624b44b0e988d0acea0b3214e694730a055cb6b2efa801"
+checksum = "62e0021ea2c22aed41653bc7e1419abb2c97e038ff2c33d0e1309e49a97deec0"
 dependencies = [
  "base64 0.22.1",
  "bytes",
@@ -5905,7 +5903,7 @@ dependencies = [
  "sync_wrapper",
  "tokio",
  "tokio-rustls",
- "tower 0.5.3",
+ "tower",
  "tower-http",
  "tower-service",
  "url",
@@ -5930,9 +5928,9 @@ dependencies = [
 
 [[package]]
 name = "roaring"
-version = "0.11.3"
+version = "0.11.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8ba9ce64a8f45d7fc86358410bb1a82e8c987504c0d4900e9141d69a9f26c885"
+checksum = "1dedc5658c6ecb3bdb5ef5f3295bb9253f42dcf3fd1402c03f6b1f7659c3c4a9"
 dependencies = [
  "bytemuck",
  "byteorder",
@@ -5950,9 +5948,9 @@ dependencies = [
 
 [[package]]
 name = "rustc-hash"
-version = "2.1.1"
+version = "2.1.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "357703d41365b4b27c590e3ed91eabb1b663f07c4c084095e60cbed4362dff0d"
+checksum = "94300abf3f1ae2e2b8ffb7b58043de3d399c73fa6f4b73826402a5c457614dbe"
 
 [[package]]
 name = "rustc_version"
@@ -5969,11 +5967,11 @@ version = "0.38.44"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "fdb5bc1ae2baa591800df16c9ca78619bf65c0488b41b96ccec5d11220d8c154"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "errno",
  "libc",
  "linux-raw-sys 0.4.15",
- "windows-sys 0.52.0",
+ "windows-sys 0.59.0",
 ]
 
 [[package]]
@@ -5982,18 +5980,18 @@ version = "1.1.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "b6fe4565b9518b83ef4f91bb47ce29620ca828bd32cb7e408f0062e9930ba190"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "errno",
  "libc",
  "linux-raw-sys 0.12.1",
- "windows-sys 0.52.0",
+ "windows-sys 0.61.2",
 ]
 
 [[package]]
 name = "rustls"
-version = "0.23.37"
+version = "0.23.40"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "758025cb5fccfd3bc2fd74708fd4682be41d99e5dff73c377c0646c6012c73a4"
+checksum = "ef86cd5876211988985292b91c96a8f2d298df24e75989a43a3c73f2d4d8168b"
 dependencies = [
  "aws-lc-rs",
  "log",
@@ -6019,9 +6017,9 @@ dependencies = [
 
 [[package]]
 name = "rustls-pki-types"
-version = "1.14.0"
+version = "1.14.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "be040f8b0a225e40375822a563fa9524378b9d63112f53e19ffff34df5d33fdd"
+checksum = "30a7197ae7eb376e574fe940d068c30fe0462554a3ddbe4eca7838e049c937a9"
 dependencies = [
  "web-time",
  "zeroize",
@@ -6029,9 +6027,9 @@ dependencies = [
 
 [[package]]
 name = "rustls-platform-verifier"
-version = "0.6.2"
+version = "0.7.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1d99feebc72bae7ab76ba994bb5e121b8d83d910ca40b36e0921f53becc41784"
+checksum = "26d1e2536ce4f35f4846aa13bff16bd0ff40157cdb14cc056c7b14ba41233ba0"
 dependencies = [
  "core-foundation 0.10.1",
  "core-foundation-sys",
@@ -6045,7 +6043,7 @@ dependencies = [
  "security-framework",
  "security-framework-sys",
  "webpki-root-certs",
- "windows-sys 0.52.0",
+ "windows-sys 0.61.2",
 ]
 
 [[package]]
@@ -6056,9 +6054,9 @@ checksum = "f87165f0995f63a9fbeea62b64d10b4d9d8e78ec6d7d51fb2125fda7bb36788f"
 
 [[package]]
 name = "rustls-webpki"
-version = "0.103.10"
+version = "0.103.13"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "df33b2b81ac578cabaf06b89b0631153a3f416b0a886e8a7a1707fb51abbd1ef"
+checksum = "61c429a8649f110dddef65e2a5ad240f747e85f7758a6bccc7e5777bd33f756e"
 dependencies = [
  "aws-lc-rs",
  "ring",
@@ -6078,7 +6076,7 @@ version = "17.0.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "e902948a25149d50edc1a8e0141aad50f54e22ba83ff988cf8f7c9ef07f50564"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "cfg-if",
  "clipboard-win",
  "fd-lock",
@@ -6176,7 +6174,7 @@ version = "3.7.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "b7f4bc775c73d9a02cde8bf7b2ec4c9d12743edf609006c7facc23998404cd1d"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "core-foundation 0.10.1",
  "core-foundation-sys",
  "libc",
@@ -6210,7 +6208,7 @@ dependencies = [
  "clap",
  "colored",
  "crossbeam-channel",
- "dashmap 6.1.0",
+ "dashmap",
  "dirs",
  "flate2",
  "futures",
@@ -6226,7 +6224,7 @@ dependencies = [
  "owo-colors",
  "rayon",
  "regex",
- "reqwest 0.13.2",
+ "reqwest 0.13.3",
  "rustyline",
  "serde",
  "serde_bytes",
@@ -6237,24 +6235,23 @@ dependencies = [
  "streaming-iterator",
  "tempfile",
  "tokio",
- "tokio-tungstenite 0.29.0",
- "tower 0.5.3",
- "tower-lsp",
+ "tokio-tungstenite",
+ "tower",
+ "tower-lsp-server",
  "tracing",
  "tracing-subscriber",
  "tree-sitter",
  "tree-sitter-c",
  "tree-sitter-python",
  "tree-sitter-rust",
- "url",
  "walkdir",
 ]
 
 [[package]]
 name = "semver"
-version = "1.0.27"
+version = "1.0.28"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d767eb0aabc880b29956c35734170f26ed551a859dbd361d140cdbeca61ab1e2"
+checksum = "8a7852d02fc848982e0c167ef163aaff9cd91dc640ba85e263cb1ce46fae51cd"
 
 [[package]]
 name = "seq-macro"
@@ -6308,7 +6305,7 @@ version = "1.0.149"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "83fc039473c5595ace860d8c4fafa220ff474b3fc6bfdb4293327f1a37e94d86"
 dependencies = [
- "indexmap 2.13.0",
+ "indexmap 2.14.0",
  "itoa",
  "memchr",
  "serde",
@@ -6352,15 +6349,15 @@ dependencies = [
 
 [[package]]
 name = "serde_with"
-version = "3.18.0"
+version = "3.19.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "dd5414fad8e6907dbdd5bc441a50ae8d6e26151a03b1de04d89a5576de61d01f"
+checksum = "f05839ce67618e14a09b286535c0d9c94e85ef25469b0e13cb4f844e5593eb19"
 dependencies = [
  "base64 0.22.1",
  "chrono",
  "hex",
  "indexmap 1.9.3",
- "indexmap 2.13.0",
+ "indexmap 2.14.0",
  "schemars 0.9.0",
  "schemars 1.2.1",
  "serde_core",
@@ -6371,9 +6368,9 @@ dependencies = [
 
 [[package]]
 name = "serde_with_macros"
-version = "3.18.0"
+version = "3.19.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d3db8978e608f1fe7357e211969fd9abdcae80bac1ba7a3369bb7eb6b404eb65"
+checksum = "cf2ebbe86054f9b45bc3881e865683ccfaccce97b9b4cb53f3039d67f355a334"
 dependencies = [
  "darling 0.23.0",
  "proc-macro2",
@@ -6388,7 +6385,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "e3bf829a2d51ab4a5ddf1352d8470c140cadc8301b2ae1789db023f01cedd6ba"
 dependencies = [
  "cfg-if",
- "cpufeatures",
+ "cpufeatures 0.2.17",
  "digest",
 ]
 
@@ -6409,7 +6406,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "a7507d819769d01a365ab707794a4084392c824f54a7a6a7862f8c3d0892b283"
 dependencies = [
  "cfg-if",
- "cpufeatures",
+ "cpufeatures 0.2.17",
  "digest",
 ]
 
@@ -6436,9 +6433,9 @@ checksum = "0fda2ff0d084019ba4d7c6f371c95d8fd75ce3524c3cb8fb653a3023f6323e64"
 
 [[package]]
 name = "signal-hook"
-version = "0.4.3"
+version = "0.4.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3b57709da74f9ff9f4a27dce9526eec25ca8407c45a7887243b031a58935fb8e"
+checksum = "b2a0c28ca5908dbdbcd52e6fdaa00358ab88637f8ab33e1f188dd510eb44b53d"
 dependencies = [
  "libc",
  "signal-hook-registry",
@@ -6456,9 +6453,19 @@ dependencies = [
 
 [[package]]
 name = "simd-adler32"
-version = "0.3.8"
+version = "0.3.9"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e320a6c5ad31d271ad523dcf3ad13e2767ad8b1cb8f047f75a8aeaf8da139da2"
+checksum = "703d5c7ef118737c72f1af64ad2f6f8c5e1921f818cdcb97b8fe6fc69bf66214"
+
+[[package]]
+name = "simd_cesu8"
+version = "1.1.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "94f90157bb87cddf702797c5dadfa0be7d266cdf49e22da2fcaa32eff75b2c33"
+dependencies = [
+ "rustc_version",
+ "simdutf8",
+]
 
 [[package]]
 name = "simdutf8"
@@ -6474,9 +6481,9 @@ checksum = "bbbb5d9659141646ae647b42fe094daf6c6192d1620870b449d9557f748b2daa"
 
 [[package]]
 name = "siphasher"
-version = "1.0.2"
+version = "1.0.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b2aa850e253778c88a04c3d7323b043aeda9d3e30d5971937c1855769763678e"
+checksum = "8ee5873ec9cce0195efcb7a4e9507a04cd49aec9c83d0389df45b1ef7ba2e649"
 
 [[package]]
 name = "sketches-ddsketch"
@@ -6548,7 +6555,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "3a766e1110788c36f4fa1c2b71b387a7815aa65f88ce0229841826633d93723e"
 dependencies = [
  "libc",
- "windows-sys 0.60.2",
+ "windows-sys 0.61.2",
 ]
 
 [[package]]
@@ -6726,7 +6733,7 @@ version = "0.7.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "a13f3d0daba03132c0aa9767f98351b3488edc2c100cda2d2ec2b04f3d8d3c8b"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "core-foundation 0.9.4",
  "system-configuration-sys",
 ]
@@ -6793,7 +6800,7 @@ dependencies = [
  "tantivy-stacker",
  "tantivy-tokenizer-api",
  "tempfile",
- "thiserror 2.0.18",
+ "thiserror",
  "time",
  "uuid",
  "winapi",
@@ -6909,16 +6916,7 @@ dependencies = [
  "getrandom 0.4.2",
  "once_cell",
  "rustix 1.1.4",
- "windows-sys 0.52.0",
-]
-
-[[package]]
-name = "thiserror"
-version = "1.0.69"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b6aaf5339b578ea85b50e080feb250a3e8ae8cfcdff9a461c9ec2904bc923f52"
-dependencies = [
- "thiserror-impl 1.0.69",
+ "windows-sys 0.61.2",
 ]
 
 [[package]]
@@ -6927,18 +6925,7 @@ version = "2.0.18"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "4288b5bcbc7920c07a1149a35cf9590a2aa808e0bc1eafaade0b80947865fbc4"
 dependencies = [
- "thiserror-impl 2.0.18",
-]
-
-[[package]]
-name = "thiserror-impl"
-version = "1.0.69"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "4fee6c4efc90059e10f81e6d42c60a18f76588c3d74cb83a0b242a2b6c7504c1"
-dependencies = [
- "proc-macro2",
- "quote",
- "syn 2.0.117",
+ "thiserror-impl",
 ]
 
 [[package]]
@@ -7012,9 +6999,9 @@ dependencies = [
 
 [[package]]
 name = "tinystr"
-version = "0.8.2"
+version = "0.8.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "42d3e9c45c09de15d06dd8acf5f4e0e399e85927b7f00711024eb7ae10fa4869"
+checksum = "c8323304221c2a851516f22236c5722a72eaa19749016521d6dff0824447d96d"
 dependencies = [
  "displaydoc",
  "zerovec",
@@ -7056,7 +7043,7 @@ dependencies = [
  "monostate",
  "onig",
  "paste",
- "rand 0.9.2",
+ "rand 0.9.4",
  "rayon",
  "rayon-cond",
  "regex",
@@ -7064,7 +7051,7 @@ dependencies = [
  "serde",
  "serde_json",
  "spm_precompiled",
- "thiserror 2.0.18",
+ "thiserror",
  "unicode-normalization-alignments",
  "unicode-segmentation",
  "unicode_categories",
@@ -7072,9 +7059,9 @@ dependencies = [
 
 [[package]]
 name = "tokio"
-version = "1.50.0"
+version = "1.52.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "27ad5e34374e03cfffefc301becb44e9dc3c17584f414349ebe29ed26661822d"
+checksum = "110a78583f19d5cdb2c5ccf321d1290344e71313c6c37d43520d386027d18386"
 dependencies = [
  "bytes",
  "libc",
@@ -7089,9 +7076,9 @@ dependencies = [
 
 [[package]]
 name = "tokio-macros"
-version = "2.6.1"
+version = "2.7.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "5c55a2eff8b69ce66c84f85e1da1c233edc36ceb85a2058d11b0d6a3c7e7569c"
+checksum = "385a6cb71ab9ab790c5fe8d67f1645e6c450a7ce006a33de03daa956cf70a496"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -7119,18 +7106,6 @@ dependencies = [
  "tokio",
 ]
 
-[[package]]
-name = "tokio-tungstenite"
-version = "0.28.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d25a406cddcc431a75d3d9afc6a7c0f7428d4891dd973e4d54c56b46127bf857"
-dependencies = [
- "futures-util",
- "log",
- "tokio",
- "tungstenite 0.28.0",
-]
-
 [[package]]
 name = "tokio-tungstenite"
 version = "0.29.0"
@@ -7140,7 +7115,7 @@ dependencies = [
  "futures-util",
  "log",
  "tokio",
- "tungstenite 0.29.0",
+ "tungstenite",
 ]
 
 [[package]]
@@ -7156,20 +7131,6 @@ dependencies = [
  "tokio",
 ]
 
-[[package]]
-name = "tower"
-version = "0.4.13"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b8fa9be0de6cf49e536ce1851f987bd21a43b771b09473c3549a6c853db37c1c"
-dependencies = [
- "futures-core",
- "futures-util",
- "pin-project",
- "pin-project-lite",
- "tower-layer",
- "tower-service",
-]
-
 [[package]]
 name = "tower"
 version = "0.5.3"
@@ -7193,7 +7154,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "d4e6559d53cc268e5031cd8429d05415bc4cb4aefc4aa5d6cc35fbf5b924a1f8"
 dependencies = [
  "async-compression",
- "bitflags 2.11.0",
+ "bitflags",
  "bytes",
  "futures-core",
  "futures-util",
@@ -7204,7 +7165,7 @@ dependencies = [
  "pin-project-lite",
  "tokio",
  "tokio-util",
- "tower 0.5.3",
+ "tower",
  "tower-layer",
  "tower-service",
 ]
@@ -7216,39 +7177,25 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "121c2a6cda46980bb0fcd1647ffaf6cd3fc79a013de288782836f6df9c48780e"
 
 [[package]]
-name = "tower-lsp"
-version = "0.20.0"
+name = "tower-lsp-server"
+version = "0.23.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d4ba052b54a6627628d9b3c34c176e7eda8359b7da9acd497b9f20998d118508"
+checksum = "2f0e711655c89181a6bc6a2cc348131fcd9680085f5b06b6af13427a393a6e72"
 dependencies = [
- "async-trait",
- "auto_impl",
  "bytes",
- "dashmap 5.5.3",
+ "dashmap",
  "futures",
  "httparse",
- "lsp-types",
+ "ls-types",
  "memchr",
  "serde",
  "serde_json",
  "tokio",
  "tokio-util",
- "tower 0.4.13",
- "tower-lsp-macros",
+ "tower",
  "tracing",
 ]
 
-[[package]]
-name = "tower-lsp-macros"
-version = "0.9.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "84fd902d4e0b9a4b27f2f440108dc034e1758628a9b702f8ec61ad66355422fa"
-dependencies = [
- "proc-macro2",
- "quote",
- "syn 2.0.117",
-]
-
 [[package]]
 name = "tower-service"
 version = "0.3.3"
@@ -7319,9 +7266,9 @@ dependencies = [
 
 [[package]]
 name = "tree-sitter"
-version = "0.26.7"
+version = "0.26.8"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e7a6592b1aec0109df37b6bafea77eb4e61466e37b0a5a98bef4f89bfb81b7a2"
+checksum = "887bd495d0582c5e3e0d8ece2233666169fa56a9644d172fc22ad179ab2d0538"
 dependencies = [
  "cc",
  "regex",
@@ -7333,9 +7280,9 @@ dependencies = [
 
 [[package]]
 name = "tree-sitter-c"
-version = "0.24.1"
+version = "0.24.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1a3aad8f0129083a59fe8596157552d2bb7148c492d44c21558d68ca1c722707"
+checksum = "a9b2eb57a55fed6b00812912e730b7a275cf4fe98bfd6a5d76263d4438371728"
 dependencies = [
  "cc",
  "tree-sitter-language",
@@ -7359,9 +7306,9 @@ dependencies = [
 
 [[package]]
 name = "tree-sitter-rust"
-version = "0.24.1"
+version = "0.24.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f715f73a0687261ddb686f0d64a1e5af57bd199c4d12be5fdda6676ce1885bf9"
+checksum = "439e577dbe07423ec2582ac62c7531120dbfccfa6e5f92406f93dd271a120e45"
 dependencies = [
  "cc",
  "tree-sitter-language",
@@ -7373,23 +7320,6 @@ version = "0.2.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "e421abadd41a4225275504ea4d6566923418b7f05506fbc9c0fe86ba7396114b"
 
-[[package]]
-name = "tungstenite"
-version = "0.28.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8628dcc84e5a09eb3d8423d6cb682965dea9133204e8fb3efee74c2a0c259442"
-dependencies = [
- "bytes",
- "data-encoding",
- "http",
- "httparse",
- "log",
- "rand 0.9.2",
- "sha1",
- "thiserror 2.0.18",
- "utf-8",
-]
-
 [[package]]
 name = "tungstenite"
 version = "0.29.0"
@@ -7401,9 +7331,9 @@ dependencies = [
  "http",
  "httparse",
  "log",
- "rand 0.9.2",
+ "rand 0.9.4",
  "sha1",
- "thiserror 2.0.18",
+ "thiserror",
 ]
 
 [[package]]
@@ -7412,14 +7342,14 @@ version = "2.1.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "9ea3136b675547379c4bd395ca6b938e5ad3c3d20fad76e7fe85f9e0d011419c"
 dependencies = [
- "rand 0.9.2",
+ "rand 0.9.4",
 ]
 
 [[package]]
 name = "typenum"
-version = "1.19.0"
+version = "1.20.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "562d481066bde0658276a35467c4af00bdc6ee726305698a55b86e61d7ad82bb"
+checksum = "40ce102ab67701b8526c123c1bab5cbe42d7040ccfd0f64af1a385808d2f43de"
 
 [[package]]
 name = "uluru"
@@ -7468,9 +7398,9 @@ dependencies = [
 
 [[package]]
 name = "unicode-segmentation"
-version = "1.12.0"
+version = "1.13.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f6ccf251212114b54433ec949fd6a7841275f9ada20dddd2f29e9ceea4501493"
+checksum = "9629274872b2bfaf8d66f5f15725007f635594914870f65218920345aa11aa8c"
 
 [[package]]
 name = "unicode-width"
@@ -7531,15 +7461,8 @@ dependencies = [
  "idna",
  "percent-encoding",
  "serde",
- "serde_derive",
 ]
 
-[[package]]
-name = "utf-8"
-version = "0.7.6"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "09cc8ee72d2a9becf2f2febe0205bbed8fc6615b7cb429ad062dc7b7ddd036a9"
-
 [[package]]
 name = "utf8-ranges"
 version = "1.0.5"
@@ -7560,9 +7483,9 @@ checksum = "06abde3611657adf66d383f00b093d7faecc7fa57071cce2578660c9f1010821"
 
 [[package]]
 name = "uuid"
-version = "1.22.0"
+version = "1.23.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a68d3c8f01c0cfa54a75291d83601161799e4a89a39e0929f4b0354d88757a37"
+checksum = "ddd74a9687298c6858e9b88ec8935ec45d22e8fd5e6394fa1bd4e99a87789c76"
 dependencies = [
  "getrandom 0.4.2",
  "js-sys",
@@ -7615,11 +7538,11 @@ checksum = "ccf3ec651a847eb01de73ccad15eb7d99f80485de043efb2f370cd654f4ea44b"
 
 [[package]]
 name = "wasip2"
-version = "1.0.2+wasi-0.2.9"
+version = "1.0.3+wasi-0.2.9"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9517f9239f02c069db75e65f174b3da828fe5f5b945c4dd26bd25d89c03ebcf5"
+checksum = "20064672db26d7cdc89c7798c48a0fdfac8213434a1186e5ef29fd560ae223d6"
 dependencies = [
- "wit-bindgen",
+ "wit-bindgen 0.57.1",
 ]
 
 [[package]]
@@ -7628,14 +7551,14 @@ version = "0.4.0+wasi-0.3.0-rc-2026-01-06"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "5428f8bf88ea5ddc08faddef2ac4a67e390b88186c703ce6dbd955e1c145aca5"
 dependencies = [
- "wit-bindgen",
+ "wit-bindgen 0.51.0",
 ]
 
 [[package]]
 name = "wasm-bindgen"
-version = "0.2.114"
+version = "0.2.120"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6532f9a5c1ece3798cb1c2cfdba640b9b3ba884f5db45973a6f442510a87d38e"
+checksum = "df52b6d9b87e0c74c9edfa1eb2d9bf85e5d63515474513aa50fa181b3c4f5db1"
 dependencies = [
  "cfg-if",
  "once_cell",
@@ -7646,23 +7569,19 @@ dependencies = [
 
 [[package]]
 name = "wasm-bindgen-futures"
-version = "0.4.64"
+version = "0.4.70"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e9c5522b3a28661442748e09d40924dfb9ca614b21c00d3fd135720e48b67db8"
+checksum = "af934872acec734c2d80e6617bbb5ff4f12b052dd8e6332b0817bce889516084"
 dependencies = [
- "cfg-if",
- "futures-util",
  "js-sys",
- "once_cell",
  "wasm-bindgen",
- "web-sys",
 ]
 
 [[package]]
 name = "wasm-bindgen-macro"
-version = "0.2.114"
+version = "0.2.120"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "18a2d50fcf105fb33bb15f00e7a77b772945a2ee45dcf454961fd843e74c18e6"
+checksum = "78b1041f495fb322e64aca85f5756b2172e35cd459376e67f2a6c9dffcedb103"
 dependencies = [
  "quote",
  "wasm-bindgen-macro-support",
@@ -7670,9 +7589,9 @@ dependencies = [
 
 [[package]]
 name = "wasm-bindgen-macro-support"
-version = "0.2.114"
+version = "0.2.120"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "03ce4caeaac547cdf713d280eda22a730824dd11e6b8c3ca9e42247b25c631e3"
+checksum = "9dcd0ff20416988a18ac686d4d4d0f6aae9ebf08a389ff5d29012b05af2a1b41"
 dependencies = [
  "bumpalo",
  "proc-macro2",
@@ -7683,9 +7602,9 @@ dependencies = [
 
 [[package]]
 name = "wasm-bindgen-shared"
-version = "0.2.114"
+version = "0.2.120"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "75a326b8c223ee17883a4251907455a2431acc2791c98c26279376490c378c16"
+checksum = "49757b3c82ebf16c57d69365a142940b384176c24df52a087fb748e2085359ea"
 dependencies = [
  "unicode-ident",
 ]
@@ -7707,7 +7626,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "bb0e353e6a2fbdc176932bbaab493762eb1255a7900fe0fea1a2f96c296cc909"
 dependencies = [
  "anyhow",
- "indexmap 2.13.0",
+ "indexmap 2.14.0",
  "wasm-encoder",
  "wasmparser",
 ]
@@ -7731,17 +7650,17 @@ version = "0.244.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "47b807c72e1bac69382b3a6fb3dbe8ea4c0ed87ff5629b8685ae6b9a611028fe"
 dependencies = [
- "bitflags 2.11.0",
+ "bitflags",
  "hashbrown 0.15.5",
- "indexmap 2.13.0",
+ "indexmap 2.14.0",
  "semver",
 ]
 
 [[package]]
 name = "web-sys"
-version = "0.3.91"
+version = "0.3.97"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "854ba17bb104abfb26ba36da9729addc7ce7f06f5c0f90f3c391f8461cca21f9"
+checksum = "2eadbac71025cd7b0834f20d1fe8472e8495821b4e9801eb0a60bd1f19827602"
 dependencies = [
  "js-sys",
  "wasm-bindgen",
@@ -7759,9 +7678,9 @@ dependencies = [
 
 [[package]]
 name = "webpki-root-certs"
-version = "1.0.6"
+version = "1.0.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "804f18a4ac2676ffb4e8b5b5fa9ae38af06df08162314f96a68d2a363e21a8ca"
+checksum = "f31141ce3fc3e300ae89b78c0dd67f9708061d1d2eda54b8209346fd6be9a92c"
 dependencies = [
  "rustls-pki-types",
 ]
@@ -7772,14 +7691,14 @@ version = "0.26.11"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "521bc38abb08001b01866da9f51eb7c5d647a19260e00054a8c7fd5f9e57f7a9"
 dependencies = [
- "webpki-roots 1.0.6",
+ "webpki-roots 1.0.7",
 ]
 
 [[package]]
 name = "webpki-roots"
-version = "1.0.6"
+version = "1.0.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "22cfaf3c063993ff62e73cb4311efde4db1efb31ab78a3e5c457939ad5cc0bed"
+checksum = "52f5ee44c96cf55f1b349600768e3ece3a8f26010c05265ab73f945bb1a2eb9d"
 dependencies = [
  "rustls-pki-types",
 ]
@@ -7806,7 +7725,7 @@ version = "0.1.11"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "c2a7b1c03c876122aa43f3020e6c3c3ee5c05081c9a00739faf7503aeba10d22"
 dependencies = [
- "windows-sys 0.52.0",
+ "windows-sys 0.61.2",
 ]
 
 [[package]]
@@ -7885,15 +7804,6 @@ dependencies = [
  "windows-link",
 ]
 
-[[package]]
-name = "windows-sys"
-version = "0.45.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "75283be5efb2831d37ea142365f009c02ec203cd29a3ebecbc093d52315b66d0"
-dependencies = [
- "windows-targets 0.42.2",
-]
-
 [[package]]
 name = "windows-sys"
 version = "0.52.0"
@@ -7930,21 +7840,6 @@ dependencies = [
  "windows-link",
 ]
 
-[[package]]
-name = "windows-targets"
-version = "0.42.2"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8e5180c00cd44c9b1c88adb3693291f1cd93605ded80c250a75d472756b4d071"
-dependencies = [
- "windows_aarch64_gnullvm 0.42.2",
- "windows_aarch64_msvc 0.42.2",
- "windows_i686_gnu 0.42.2",
- "windows_i686_msvc 0.42.2",
- "windows_x86_64_gnu 0.42.2",
- "windows_x86_64_gnullvm 0.42.2",
- "windows_x86_64_msvc 0.42.2",
-]
-
 [[package]]
 name = "windows-targets"
 version = "0.52.6"
@@ -7978,12 +7873,6 @@ dependencies = [
  "windows_x86_64_msvc 0.53.1",
 ]
 
-[[package]]
-name = "windows_aarch64_gnullvm"
-version = "0.42.2"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "597a5118570b68bc08d8d59125332c54f1ba9d9adeedeef5b99b02ba2b0698f8"
-
 [[package]]
 name = "windows_aarch64_gnullvm"
 version = "0.52.6"
@@ -7996,12 +7885,6 @@ version = "0.53.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "a9d8416fa8b42f5c947f8482c43e7d89e73a173cead56d044f6a56104a6d1b53"
 
-[[package]]
-name = "windows_aarch64_msvc"
-version = "0.42.2"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "e08e8864a60f06ef0d0ff4ba04124db8b0fb3be5776a5cd47641e942e58c4d43"
-
 [[package]]
 name = "windows_aarch64_msvc"
 version = "0.52.6"
@@ -8014,12 +7897,6 @@ version = "0.53.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "b9d782e804c2f632e395708e99a94275910eb9100b2114651e04744e9b125006"
 
-[[package]]
-name = "windows_i686_gnu"
-version = "0.42.2"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c61d927d8da41da96a81f029489353e68739737d3beca43145c8afec9a31a84f"
-
 [[package]]
 name = "windows_i686_gnu"
 version = "0.52.6"
@@ -8044,12 +7921,6 @@ version = "0.53.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "fa7359d10048f68ab8b09fa71c3daccfb0e9b559aed648a8f95469c27057180c"
 
-[[package]]
-name = "windows_i686_msvc"
-version = "0.42.2"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "44d840b6ec649f480a41c8d80f9c65108b92d89345dd94027bfe06ac444d1060"
-
 [[package]]
 name = "windows_i686_msvc"
 version = "0.52.6"
@@ -8062,12 +7933,6 @@ version = "0.53.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "1e7ac75179f18232fe9c285163565a57ef8d3c89254a30685b57d83a38d326c2"
 
-[[package]]
-name = "windows_x86_64_gnu"
-version = "0.42.2"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8de912b8b8feb55c064867cf047dda097f92d51efad5b491dfb98f6bbb70cb36"
-
 [[package]]
 name = "windows_x86_64_gnu"
 version = "0.52.6"
@@ -8080,12 +7945,6 @@ version = "0.53.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "9c3842cdd74a865a8066ab39c8a7a473c0778a3f29370b5fd6b4b9aa7df4a499"
 
-[[package]]
-name = "windows_x86_64_gnullvm"
-version = "0.42.2"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "26d41b46a36d453748aedef1486d5c7a85db22e56aff34643984ea85514e94a3"
-
 [[package]]
 name = "windows_x86_64_gnullvm"
 version = "0.52.6"
@@ -8098,12 +7957,6 @@ version = "0.53.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "0ffa179e2d07eee8ad8f57493436566c7cc30ac536a3379fdf008f47f6bb7ae1"
 
-[[package]]
-name = "windows_x86_64_msvc"
-version = "0.42.2"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9aec5da331524158c6d1a4ac0ab1541149c0b9505fde06423b02f5ef0106b9f0"
-
 [[package]]
 name = "windows_x86_64_msvc"
 version = "0.52.6"
@@ -8134,6 +7987,12 @@ dependencies = [
  "wit-bindgen-rust-macro",
 ]
 
+[[package]]
+name = "wit-bindgen"
+version = "0.57.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "1ebf944e87a7c253233ad6766e082e3cd714b5d03812acc24c318f549614536e"
+
 [[package]]
 name = "wit-bindgen-core"
 version = "0.51.0"
@@ -8153,7 +8012,7 @@ checksum = "b7c566e0f4b284dd6561c786d9cb0142da491f46a9fbed79ea69cdad5db17f21"
 dependencies = [
  "anyhow",
  "heck",
- "indexmap 2.13.0",
+ "indexmap 2.14.0",
  "prettyplease",
  "syn 2.0.117",
  "wasm-metadata",
@@ -8183,8 +8042,8 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "9d66ea20e9553b30172b5e831994e35fbde2d165325bec84fc43dbf6f4eb9cb2"
 dependencies = [
  "anyhow",
- "bitflags 2.11.0",
- "indexmap 2.13.0",
+ "bitflags",
+ "indexmap 2.14.0",
  "log",
  "serde",
  "serde_derive",
@@ -8203,7 +8062,7 @@ checksum = "ecc8ac4bc1dc3381b7f59c34f00b67e18f910c2c0f50015669dde7def656a736"
 dependencies = [
  "anyhow",
  "id-arena",
- "indexmap 2.13.0",
+ "indexmap 2.14.0",
  "log",
  "semver",
  "serde",
@@ -8215,9 +8074,9 @@ dependencies = [
 
 [[package]]
 name = "writeable"
-version = "0.6.2"
+version = "0.6.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9edde0db4769d2dc68579893f2306b26c6ecfbe0ef499b013d731b7b9247e0b9"
+checksum = "1ffae5123b2d3fc086436f8834ae3ab053a283cfac8fe0a0b8eaae044768a4c4"
 
 [[package]]
 name = "wyz"
@@ -8236,9 +8095,9 @@ checksum = "fdd20c5420375476fbd4394763288da7eb0cc0b8c11deed431a91562af7335d3"
 
 [[package]]
 name = "yoke"
-version = "0.8.1"
+version = "0.8.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "72d6e5c6afb84d73944e5cedb052c4680d5657337201555f9f2a16b7406d4954"
+checksum = "abe8c5fda708d9ca3df187cae8bfb9ceda00dd96231bed36e445a1a48e66f9ca"
 dependencies = [
  "stable_deref_trait",
  "yoke-derive",
@@ -8247,9 +8106,9 @@ dependencies = [
 
 [[package]]
 name = "yoke-derive"
-version = "0.8.1"
+version = "0.8.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b659052874eb698efe5b9e8cf382204678a0086ebf46982b79d6ca3182927e5d"
+checksum = "de844c262c8848816172cef550288e7dc6c7b7814b4ee56b3e1553f275f1858e"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -8259,18 +8118,18 @@ dependencies = [
 
 [[package]]
 name = "zerocopy"
-version = "0.8.47"
+version = "0.8.48"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "efbb2a062be311f2ba113ce66f697a4dc589f85e78a4aea276200804cea0ed87"
+checksum = "eed437bf9d6692032087e337407a86f04cd8d6a16a37199ed57949d415bd68e9"
 dependencies = [
  "zerocopy-derive",
 ]
 
 [[package]]
 name = "zerocopy-derive"
-version = "0.8.47"
+version = "0.8.48"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "0e8bc7269b54418e7aeeef514aa68f8690b8c0489a06b0136e5f57c4c5ccab89"
+checksum = "70e3cd084b1788766f53af483dd21f93881ff30d7320490ec3ef7526d203bad4"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -8279,18 +8138,18 @@ dependencies = [
 
 [[package]]
 name = "zerofrom"
-version = "0.1.6"
+version = "0.1.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "50cc42e0333e05660c3587f3bf9d0478688e15d870fab3346451ce7f8c9fbea5"
+checksum = "69faa1f2a1ea75661980b013019ed6687ed0e83d069bc1114e2cc74c6c04c4df"
 dependencies = [
  "zerofrom-derive",
 ]
 
 [[package]]
 name = "zerofrom-derive"
-version = "0.1.6"
+version = "0.1.7"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d71e5d6e06ab090c67b5e44993ec16b72dcbaabc526db883a360057678b48502"
+checksum = "11532158c46691caf0f2593ea8358fed6bbf68a0315e80aae9bd41fbade684a1"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -8306,9 +8165,9 @@ checksum = "b97154e67e32c85465826e8bcc1c59429aaaf107c1e4a9e53c8d8ccd5eff88d0"
 
 [[package]]
 name = "zerotrie"
-version = "0.2.3"
+version = "0.2.4"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "2a59c17a5562d507e4b54960e8569ebee33bee890c70aa3fe7b97e85a9fd7851"
+checksum = "0f9152d31db0792fa83f70fb2f83148effb5c1f5b8c7686c3459e361d9bc20bf"
 dependencies = [
  "displaydoc",
  "yoke",
@@ -8317,9 +8176,9 @@ dependencies = [
 
 [[package]]
 name = "zerovec"
-version = "0.11.5"
+version = "0.11.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "6c28719294829477f525be0186d13efa9a3c602f7ec202ca9e353d310fb9a002"
+checksum = "90f911cbc359ab6af17377d242225f4d75119aec87ea711a880987b18cd7b239"
 dependencies = [
  "yoke",
  "zerofrom",
@@ -8328,9 +8187,9 @@ dependencies = [
 
 [[package]]
 name = "zerovec-derive"
-version = "0.11.2"
+version = "0.11.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "eadce39539ca5cb3985590102671f2567e659fca9666581ad3411d59207951f3"
+checksum = "625dc425cab0dca6dc3c3319506e6593dcb08a9f387ea3b284dbd52a92c40555"
 dependencies = [
  "proc-macro2",
  "quote",
diff --git a/Cargo.toml b/Cargo.toml
index fd3ce2f..8f03437 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -17,7 +17,7 @@ tokio = { version = "1.50.0", features = ["full"] }
 serde = { version = "1.0.228", features = ["derive"] }
 serde_json = { version = "1.0", features = ["preserve_order"] }
 serde_bytes = "0.11"
-clap = { version = "4.6", features = ["derive"] }
+clap = { version = "4.6", features = ["derive", "env"] }
 anyhow = "1.0"
 walkdir = "2.5"
 indicatif = "0.18.4"
@@ -51,8 +51,7 @@ anstream = "1.0"        # Auto TTY/NO_COLOR handling; print!/println! and Write
 owo-colors = { version = "4", features = ["supports-colors"] }  # .red().bold(), optional detection helpers [web:27]
 gix = { version = "0.81", features = ["blocking-network-client", "worktree-mutation", "blocking-http-transport-curl-openssl"] }
 model2vec-rs = "0.1.4"
-tower-lsp = "0.20"      # Language Server Protocol framework for LSP server
-url = "2.5"             # URL parsing for file URIs in LSP
+tower-lsp-server = "0.23"  # Language Server Protocol framework for LSP server
 similar = { version = "2.7", features = ["text"] }  # Proper diff algorithm for commit diffs
 shlex = "1.3"           # Shell-like parsing for quoted strings in CLI
 reqwest = { version = "0.13", features = ["blocking", "gzip"] }  # HTTP client for manifest downloads
diff --git a/docs/schema.md b/docs/schema.md
index 963b005..8b9fb9c 100644
--- a/docs/schema.md
+++ b/docs/schema.md
@@ -363,6 +363,7 @@ branch_name         (Utf8, NOT NULL)     - Branch name (e.g., "main", "origin/de
 tip_commit          (Utf8, NOT NULL)     - Commit SHA at the tip when indexed (40-char hex)
 indexed_at          (Int64, NOT NULL)    - Unix timestamp of when branch was last indexed
 remote              (Utf8, nullable)     - Remote name if tracking branch (e.g., "origin")
+repo_path           (Utf8, nullable)     - Canonical repository identity path for multi-repo scoping
 ```
 
 **Purpose:**
@@ -370,6 +371,7 @@ remote              (Utf8, nullable)     - Remote name if tracking branch (e.g.,
 - Enables efficient multi-branch indexing by skipping already-current branches
 - Supports both local branches (e.g., "main") and remote-tracking branches (e.g., "origin/develop")
 - Stores indexing timestamp for freshness tracking
+- Scopes branch references per repository in multi-repository shared database mode
 
 **Use Cases:**
 - Multi-branch indexing: `semcode-index --branches main,develop,feature-x`
@@ -381,6 +383,7 @@ remote              (Utf8, nullable)     - Remote name if tracking branch (e.g.,
 - BTree on `branch_name` (primary lookup by branch name)
 - BTree on `tip_commit` (find branches at specific commits)
 - BTree on `remote` (filter by remote)
+- BTree on `repo_path` (fast repo-scoped branch query pruning)
 
 ---
 
@@ -509,10 +512,20 @@ The git_commits table stores git commit history with enhanced metadata:
 - **Walk-back Symbol Extraction**: Fast O(modified_lines × 50) algorithm identifies changed functions, types, and macros
 - **Dual-file Analysis**: Extracts symbols from both additions and deletions
 - **Enhanced Hunk Headers**: Git-style `@@ ... @@ symbol` format for better context
-- **Commit Traversal**: Parent relationships enable git history analysis
-- **Tag Parsing**: Extracts structured metadata from commit messages (Signed-off-by, Reviewed-by, etc.)
 - **Use Cases**: Commit analysis, code evolution tracking, review assistance, git history search
 
+### Multi-Repository Shared Database Support & Database Locality
+
+**Overview:**
+When operating `semcode` with a shared central database (e.g., `SEMCODE_DB=~/.cache/semcode.db` or an external directory shared across multiple trees of a project), branch records are isolated by repository identity (`repo_path`) to prevent cross-repository branch pollution.
+
+**Database Locality Modes:**
+- **Local Database Mode**: When the database resides inside a repository directory (such as `.git/` or `.semcode.db`), branch store queries use the fallback filter `(repo_path = '{repo_identity}' OR repo_path IS NULL)`. This ensures position-independent relocatability and full backward compatibility with unscoped legacy records while matching current repository branches.
+- **Shared / Central Database Mode**: When the database resides in a central location outside all repositories, queries strictly filter `repo_path = '{repo_identity}'` to ensure clean separation between repositories.
+- **Worktree Identity Consolidation**: Linked Git worktrees automatically resolve to their primary repository identity (`common_dir`), sharing the same branch index without duplication.
+- **On-Write Lazy Migration**: Older database tables are lazily migrated on write operations (`record_branch_indexed`) by adding the `repo_path` column JIT without interrupting read queries.
+- **Automatic Garbage Collection**: `DatabaseManager::garbage_collect()` automatically purges branch records corresponding to deleted filesystem paths in shared mode, and purges legacy unscoped `NULL` records when running in standalone database mode.
+
 ## Query Patterns
 
 ### Basic Lookups
diff --git a/docs/shared-db.md b/docs/shared-db.md
new file mode 100644
index 0000000..ef1fd24
--- /dev/null
+++ b/docs/shared-db.md
@@ -0,0 +1,104 @@
+# Semcode Shared Database Architecture & Usage Guide
+
+`semcode` supports operating with a **Shared Central Database** shared across multiple repository trees, subsystem forks, and worktrees of the same project (such as Linux kernel trees: `linux-mainline`, `bpf-next`, `net-next`, `stable`).
+
+This document explains why and how Shared Database mode works, how cross-repository data deduplication and branch isolation are maintained, and how to use it.
+
+---
+
+## 1. Why Use Shared Database Mode?
+
+When working on large projects like the Linux kernel, developers frequently maintain multiple repository checkouts:
+- Mainline (`linux-mainline`)
+- Subsystem trees (`bpf-next`, `net-next`, `sound`)
+- Stable release trees (`linux-6.12.y`, `linux-6.1.y`)
+- Multiple local Git worktrees (`git worktree add ...`)
+
+Because these trees share 90%+ of their Git commit graph and source code, indexing each tree into a separate database creates massive data duplication and wastes indexing time.
+
+### Key Benefits of Shared Database Mode:
+1. **Near-Instant Incremental Indexing Across Trees**: When you index a new subsystem tree or worktree into a shared database, `semcode` reuses all existing commit objects, file parsing results, and symbol AST definitions already indexed from other trees.
+2. **Dramatic Storage Savings (50-80%)**: Content-addressed tables (functions, types, commit diffs, Blake3 content shards) are globally deduplicated across all repositories.
+3. **Clean Branch Isolation**: Branch references (`indexed_branches`) are scoped per repository tree, preventing branch name collisions (e.g. `main` or `v6.12` existing in different trees).
+4. **Git Worktree Consolidation**: Linked worktrees automatically map to their primary repository identity (`common_dir`), sharing branch references seamlessly without duplicate worktree records.
+
+---
+
+## 2. Core Architecture
+
+### Global Content Deduplication vs. Relational Scoping
+
+`semcode` divides its database tables into two distinct categories:
+
+#### 1. Global Content Tables
+* **Tables**: `functions`, `types`, `processed_files`, `git_commits`, `commit_vectors`, `content_0..15`
+* **Strategy**: **Globally Deduplicated**
+* **Details**: Keyed on SHA-1 file blob hashes, Blake3 content hashes, or Git commit SHAs. Identical function definitions or commit diffs across different repository trees are stored **once**.
+
+#### 2. Scoped Identity Tables
+* **Tables**: `indexed_branches`
+* **Strategy**: **Repo-Scoped**
+* **Details**: Stores `repo_path` (canonical repository identity path) to isolate branch references (`main`, `master`, `v6.12`) per repository checkout.
+
+---
+
+## 3. Database Locality Modes (`Local` vs `Shared`)
+
+`semcode` automatically detects the database locality mode based on where the database file resides:
+
+### 1. Local Database Mode (`.semcode.db`)
+* **Trigger**: The database file is located inside a repository directory (e.g., `/home/user/linux/.semcode.db` or `/home/user/linux/.git/semcode.db`).
+* **Behavior**:
+  * Records `repo_path = NULL` or matches dynamically using `(repo_path = '{repo_identity}' OR repo_path IS NULL)`.
+  * **100% Relocatable**: Moving or renaming the repository directory on disk will **never** invalidate database records or break future queries.
+
+### 2. Shared Database Mode (`SEMCODE_DB`)
+* **Trigger**: The database file resides in a central directory outside all repositories (e.g., `SEMCODE_DB=~/.cache/semcode.db` or `/var/shared/semcode.db`).
+* **Behavior**:
+  * Enforces strict SQL filtering on branch queries: `repo_path = '{repo_identity}'`.
+  * Ensures complete isolation between different repository checkouts.
+
+---
+
+## 4. Git Worktree Identity Consolidation
+
+Linked Git worktrees created via `git worktree add` share the same `.git` commit graph and refs as the parent repository. 
+
+`semcode` automatically resolves worktrees to their parent repository's canonical identity (`common_dir`):
+* Primary Checkout (`/home/user/linux`): Identity path `~/linux/.git`
+* Linked Worktree (`/home/user/linux-feature-wt`): Identity path `~/linux/.git`
+
+Because both resolve to `~/linux/.git`, worktrees transparently share the same branch index without creating redundant branch records.
+
+---
+
+## 5. Automatic Garbage Collection
+
+In Shared Database mode, deleting a repository directory from the filesystem leaves behind stale `indexed_branches` records for that path.
+
+`semcode-index` automatically invokes `DatabaseManager::garbage_collect()` at the end of every indexing run:
+* **In Shared DB Mode**: Scans `indexed_branches` for `repo_path` entries whose paths no longer exist on the local filesystem and purges them via `table.delete()`.
+* **In Local DB Mode**: Purges legacy unscoped `NULL` branch records if the database is running standalone outside of Git repositories.
+
+---
+
+## 6. How to Use Shared Database Mode
+
+To use a shared central database across all your projects, set the `SEMCODE_DB` environment variable:
+
+```bash
+# 1. Index your mainline checkout into the central database
+export SEMCODE_DB=~/.cache/semcode.db
+cd ~/projects/linux-mainline
+semcode-index -s .
+
+# 2. Index a subsystem fork into the same central database (runs near-instantly!)
+cd ~/projects/bpf-next
+semcode-index -s .
+
+# 3. Query within your current checkout (queries are automatically scoped to the active repository you are in)
+cd ~/projects/bpf-next
+semcode find_function bpf_prog_load
+```
+
+You can also add `export SEMCODE_DB=~/.cache/semcode.db` to your `~/.bashrc` or `~/.zshrc` to make all `semcode`, `semcode-index`, and MCP server instances automatically use the shared central database!
diff --git a/scripts/nomic2vec.py b/scripts/nomic2vec.py
index 0025d4c..1977ab5 100755
--- a/scripts/nomic2vec.py
+++ b/scripts/nomic2vec.py
@@ -493,6 +493,17 @@ def tokens(self, value):
                 # Allow setting but always return self
                 pass
 
+        # NomicBertEmbedding doesn't implement {get,set}_input_embeddings, but
+        # it is required for distill_from_model. See:
+        # https://huggingface.co/nomic-ai/nomic-bert-2048/discussions/22
+        import types
+        def get_input_embeddings(self):
+            return self.embeddings.word_embeddings
+        def set_input_embeddings(self, value):
+            self.embeddings.word_embeddings = value
+        model.get_input_embeddings = types.MethodType(get_input_embeddings, model)
+        model.set_input_embeddings = types.MethodType(set_input_embeddings, model)
+
         try:
             # Get full stack trace to see where the error is coming from
             import traceback
diff --git a/scripts/requirements.txt b/scripts/requirements.txt
index 675faf6..68b50f0 100644
--- a/scripts/requirements.txt
+++ b/scripts/requirements.txt
@@ -8,7 +8,7 @@ tokenizers>=0.13.0
 numpy>=1.24.0
 requests>=2.25.0
 tqdm>=4.60.0
-model2vec>=0.1.0
+model2vec[distill]>=0.1.0
 sentence-transformers>=2.2.0
 sentencepiece>=0.1.97
 einops>=0.8.1
diff --git a/src/bin/index.rs b/src/bin/index.rs
index f1f8f17..c5995b1 100644
--- a/src/bin/index.rs
+++ b/src/bin/index.rs
@@ -1080,30 +1080,30 @@ async fn main() -> Result<()> {
                 total_emails_all_archives
             );
 
-            // Optimize before creating FTS indices so that compaction
-            // does not orphan the index data that was just built.
             if total_new_emails > 0 {
-                match db_manager.check_optimization_health().await {
-                    Ok((needs_optimization, message)) => {
-                        if needs_optimization {
-                            println!("\n{}", message);
-                            match db_manager.optimize_database().await {
-                                Ok(_) => println!("Database optimization completed successfully"),
-                                Err(e) => error!("Failed to optimize database: {}", e),
-                            }
-                        } else {
-                            println!("\n{}", message);
-                        }
-                    }
-                    Err(e) => {
-                        error!("Failed to check database health: {}", e);
-                    }
+                // Compact only lore tables.  The full optimize_database()
+                // method processes every table in the database including
+                // code-index tables and content shards that a lore run
+                // never touches.  On memory-constrained systems the
+                // combined working set triggers the OOM killer.
+                println!("\nCompacting lore tables...");
+                match db_manager.compact_lore_tables().await {
+                    Ok(_) => println!("Lore table compaction completed successfully"),
+                    Err(e) => error!("Failed to compact lore tables: {}", e),
                 }
 
-                println!("\nCreating FTS indices for lore table...");
-                match db_manager.create_lore_fts_indices().await {
-                    Ok(_) => println!("FTS indices created successfully"),
-                    Err(e) => eprintln!("Warning: Failed to create FTS indices: {}", e),
+                // Create FTS indices on first run; merge new rows on
+                // subsequent runs.  LanceDB's native FTS engine serves
+                // unindexed rows via brute-force fallback, so queries
+                // remain correct before optimize completes.
+                println!("\nUpdating FTS indices for lore table...");
+                match db_manager.ensure_lore_fts_indices().await {
+                    Ok(_) => {}
+                    Err(e) => eprintln!("Warning: Failed to ensure FTS indices: {}", e),
+                }
+                match db_manager.optimize_lore_fts_indices().await {
+                    Ok(_) => println!("FTS indices updated successfully"),
+                    Err(e) => eprintln!("Warning: Failed to optimize FTS indices: {}", e),
                 }
             }
 
@@ -1248,30 +1248,21 @@ async fn main() -> Result<()> {
                 }
             }
 
-            // Optimize before creating FTS indices so that compaction
-            // does not orphan the index data that was just built.
             if total_new_emails > 0 {
-                match db_manager.check_optimization_health().await {
-                    Ok((needs_optimization, message)) => {
-                        if needs_optimization {
-                            println!("\n{}", message);
-                            match db_manager.optimize_database().await {
-                                Ok(_) => println!("Database optimization completed successfully"),
-                                Err(e) => error!("Failed to optimize database: {}", e),
-                            }
-                        } else {
-                            println!("\n{}", message);
-                        }
-                    }
-                    Err(e) => {
-                        error!("Failed to check database health: {}", e);
-                    }
+                println!("\nCompacting lore tables...");
+                match db_manager.compact_lore_tables().await {
+                    Ok(_) => println!("Lore table compaction completed successfully"),
+                    Err(e) => error!("Failed to compact lore tables: {}", e),
                 }
 
-                println!("\nCreating FTS indices for lore table...");
-                match db_manager.create_lore_fts_indices().await {
-                    Ok(_) => println!("FTS indices created successfully"),
-                    Err(e) => eprintln!("Warning: Failed to create FTS indices: {}", e),
+                println!("\nUpdating FTS indices for lore table...");
+                match db_manager.ensure_lore_fts_indices().await {
+                    Ok(_) => {}
+                    Err(e) => eprintln!("Warning: Failed to ensure FTS indices: {}", e),
+                }
+                match db_manager.optimize_lore_fts_indices().await {
+                    Ok(_) => println!("FTS indices updated successfully"),
+                    Err(e) => eprintln!("Warning: Failed to optimize FTS indices: {}", e),
                 }
             }
 
@@ -1669,7 +1660,8 @@ async fn run_pipeline(args: Args) -> Result<()> {
         }
     }
 
-    // Optimization is now handled inside process_git_range for all modes
+    // Perform database garbage collection for stale branch records
+    let _ = db_manager.garbage_collect().await;
 
     // Drop and recreate if requested
     if args.drop_recreate {
diff --git a/src/bin/query.rs b/src/bin/query.rs
index ae8a374..7f92b47 100644
--- a/src/bin/query.rs
+++ b/src/bin/query.rs
@@ -49,7 +49,7 @@ struct Args {
     database: Option<String>,
 
     /// Path to the git repository for git-aware queries
-    #[arg(long, default_value = ".")]
+    #[arg(long, env = "SEMCODE_GIT_REPO", default_value = ".")]
     git_repo: String,
 
     /// Path to local model directory (for semantic search)
@@ -554,3 +554,29 @@ async fn main() -> Result<()> {
 
     Ok(())
 }
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use clap::Parser;
+    use std::env;
+
+    // Combined into a single test so the three cases don't race on the
+    // process-wide SEMCODE_GIT_REPO env var when cargo runs tests in parallel.
+    #[test]
+    fn test_git_repo_arg_env_default() {
+        let key = "SEMCODE_GIT_REPO";
+
+        env::set_var(key, "/tmp/test_repo");
+        let args = Args::try_parse_from(["query"]).unwrap();
+        assert_eq!(args.git_repo, "/tmp/test_repo");
+
+        env::set_var(key, "/tmp/env_repo");
+        let args = Args::try_parse_from(["query", "--git-repo", "/tmp/arg_repo"]).unwrap();
+        assert_eq!(args.git_repo, "/tmp/arg_repo");
+
+        env::remove_var(key);
+        let args = Args::try_parse_from(["query"]).unwrap();
+        assert_eq!(args.git_repo, ".");
+    }
+}
diff --git a/src/bin/semcode-lsp.rs b/src/bin/semcode-lsp.rs
index d8b5c8c..69d57c4 100644
--- a/src/bin/semcode-lsp.rs
+++ b/src/bin/semcode-lsp.rs
@@ -4,10 +4,9 @@ use serde::{Deserialize, Serialize};
 use std::path::Path;
 use std::sync::Arc;
 use tokio::sync::Mutex;
-use tower_lsp::jsonrpc::Result as LspResult;
-use tower_lsp::lsp_types::*;
-use tower_lsp::{Client, LanguageServer, LspService, Server};
-use url::Url;
+use tower_lsp_server::jsonrpc::Result as LspResult;
+use tower_lsp_server::ls_types::*;
+use tower_lsp_server::{Client, LanguageServer, LspService, Server};
 
 use semcode::{database_utils, DatabaseManager};
 
@@ -33,7 +32,7 @@ impl SemcodeLspBackend {
         }
     }
 
-    async fn ensure_database_connection(&self, workspace_uri: Option<&Url>) -> Result<()> {
+    async fn ensure_database_connection(&self, workspace_uri: Option<&Uri>) -> Result<()> {
         let mut db = self.database.lock().await;
         if db.is_some() {
             return Ok(());
@@ -41,6 +40,11 @@ impl SemcodeLspBackend {
 
         // Try to determine database path and git repo path from config or workspace
         let config = self.config.lock().await;
+        let env_db = std::env::var("SEMCODE_DB")
+            .ok()
+            .map(|s| s.trim().to_string())
+            .filter(|s| !s.is_empty());
+
         let (db_path, git_repo_path) = if let Some(path) = &config.database_path {
             // Custom database path provided
             let git_repo = std::env::current_dir()
@@ -48,11 +52,18 @@ impl SemcodeLspBackend {
                 .and_then(|p| p.to_str().map(|s| s.to_string()))
                 .unwrap_or_else(|| ".".to_string());
             (path.clone(), git_repo)
+        } else if let Some(env_path) = env_db {
+            // SEMCODE_DB environment variable
+            let git_repo = std::env::current_dir()
+                .ok()
+                .and_then(|p| p.to_str().map(|s| s.to_string()))
+                .unwrap_or_else(|| ".".to_string());
+            (env_path, git_repo)
         } else if let Some(uri) = workspace_uri {
             // Use workspace directory (process_database_path will add .semcode.db)
             let workspace_path = uri
                 .to_file_path()
-                .map_err(|_| anyhow::anyhow!("Failed to convert workspace URI to file path"))?;
+                .ok_or_else(|| anyhow::anyhow!("Failed to convert workspace URI to file path"))?;
             let workspace_str = workspace_path
                 .to_str()
                 .ok_or_else(|| anyhow::anyhow!("Invalid workspace path"))?
@@ -163,7 +174,7 @@ impl SemcodeLspBackend {
         drop(repo_path_guard);
 
         // Convert absolute file path to URI
-        let file_uri = Url::from_file_path(&absolute_path).ok()?;
+        let file_uri = Uri::from_file_path(&absolute_path)?;
 
         // Create position (LSP uses 0-based line numbers)
         let position = Position {
@@ -257,7 +268,7 @@ impl SemcodeLspBackend {
             drop(repo_path_guard);
 
             // Convert to URI
-            if let Ok(file_uri) = Url::from_file_path(&absolute_path) {
+            if let Some(file_uri) = Uri::from_file_path(&absolute_path) {
                 let position = Position {
                     line: line_start.saturating_sub(1),
                     character: 0,
@@ -322,19 +333,24 @@ impl SemcodeLspBackend {
     }
 }
 
-#[tower_lsp::async_trait]
 impl LanguageServer for SemcodeLspBackend {
     async fn initialize(&self, params: InitializeParams) -> LspResult<InitializeResult> {
-        // Try to establish database connection
-        let _ = self
-            .ensure_database_connection(params.root_uri.as_ref())
-            .await;
+        // Try to establish database connection using workspace folders (preferred) or root_uri (legacy)
+        let workspace_uri = params
+            .workspace_folders
+            .as_ref()
+            .and_then(|folders| folders.first())
+            .map(|f| &f.uri);
+        #[allow(deprecated)]
+        let workspace_uri = workspace_uri.or(params.root_uri.as_ref());
+        let _ = self.ensure_database_connection(workspace_uri).await;
 
         Ok(InitializeResult {
             server_info: Some(ServerInfo {
                 name: "semcode-lsp".to_string(),
                 version: Some("0.1.0".to_string()),
             }),
+            offset_encoding: None,
             capabilities: ServerCapabilities {
                 definition_provider: Some(OneOf::Left(true)),
                 references_provider: Some(OneOf::Left(true)),
@@ -360,7 +376,11 @@ impl LanguageServer for SemcodeLspBackend {
         let position = &params.text_document_position_params.position;
 
         // Get the document text to extract the function name
-        let document_text = match std::fs::read_to_string(uri.path()) {
+        let file_path = match uri.to_file_path() {
+            Some(p) => p.into_owned(),
+            None => return Ok(None),
+        };
+        let document_text = match std::fs::read_to_string(&file_path) {
             Ok(text) => text,
             Err(_) => return Ok(None),
         };
@@ -385,7 +405,11 @@ impl LanguageServer for SemcodeLspBackend {
         let position = &params.text_document_position.position;
 
         // Get the document text to extract the function name
-        let document_text = match std::fs::read_to_string(uri.path()) {
+        let file_path = match uri.to_file_path() {
+            Some(p) => p.into_owned(),
+            None => return Ok(None),
+        };
+        let document_text = match std::fs::read_to_string(&file_path) {
             Ok(text) => text,
             Err(_) => return Ok(None),
         };
diff --git a/src/bin/semcode-mcp.rs b/src/bin/semcode-mcp.rs
index 3ca0666..c8d4079 100644
--- a/src/bin/semcode-mcp.rs
+++ b/src/bin/semcode-mcp.rs
@@ -1814,7 +1814,7 @@ struct Args {
     database: Option<String>,
 
     /// Path to the git repository for git-aware queries
-    #[arg(long, default_value = ".")]
+    #[arg(long, env = "SEMCODE_GIT_REPO", default_value = ".")]
     git_repo: String,
 
     /// Path to custom model directory (defaults to ~/.cache/semcode/models/)
@@ -5680,6 +5680,28 @@ async fn main() -> Result<()> {
 #[cfg(test)]
 mod tests {
     use super::*;
+    use clap::Parser;
+    use std::env;
+
+    // Combined into a single test so the three cases don't race on the
+    // process-wide SEMCODE_GIT_REPO env var when cargo runs tests in parallel.
+    #[test]
+    fn test_git_repo_arg_env_default() {
+        let key = "SEMCODE_GIT_REPO";
+
+        env::set_var(key, "/tmp/test_repo_mcp");
+        let args = Args::try_parse_from(["semcode-mcp"]).unwrap();
+        assert_eq!(args.git_repo, "/tmp/test_repo_mcp");
+
+        env::set_var(key, "/tmp/env_repo_mcp");
+        let args =
+            Args::try_parse_from(["semcode-mcp", "--git-repo", "/tmp/arg_repo_mcp"]).unwrap();
+        assert_eq!(args.git_repo, "/tmp/arg_repo_mcp");
+
+        env::remove_var(key);
+        let args = Args::try_parse_from(["semcode-mcp"]).unwrap();
+        assert_eq!(args.git_repo, ".");
+    }
 
     #[test]
     fn test_indexing_state_new() {
diff --git a/src/database/branches.rs b/src/database/branches.rs
index 335405f..a485740 100644
--- a/src/database/branches.rs
+++ b/src/database/branches.rs
@@ -11,6 +11,7 @@ use arrow::datatypes::{DataType, Field, Schema};
 use futures::TryStreamExt;
 use lancedb::connection::Connection;
 use lancedb::query::{ExecutableQuery, QueryBase};
+use lancedb::index::{scalar::BTreeIndexBuilder, Index as LanceIndex};
 use std::sync::Arc;
 
 /// Information about an indexed branch
@@ -24,6 +25,8 @@ pub struct IndexedBranchInfo {
     pub indexed_at: i64,
     /// Remote name if this is a remote-tracking branch (e.g., "origin")
     pub remote: Option<String>,
+    /// Repository canonical path identity (NULL for local DB, canonical path for shared DB)
+    pub repo_path: Option<String>,
 }
 
 /// JSON-serializable version of IndexedBranchInfo
@@ -33,6 +36,7 @@ pub struct IndexedBranchInfoJson {
     pub tip_commit: String,
     pub indexed_at: i64,
     pub remote: Option<String>,
+    pub repo_path: Option<String>,
 }
 
 impl From<IndexedBranchInfo> for IndexedBranchInfoJson {
@@ -42,6 +46,7 @@ impl From<IndexedBranchInfo> for IndexedBranchInfoJson {
             tip_commit: info.tip_commit,
             indexed_at: info.indexed_at,
             remote: info.remote,
+            repo_path: info.repo_path,
         }
     }
 }
@@ -53,6 +58,7 @@ impl From<IndexedBranchInfoJson> for IndexedBranchInfo {
             tip_commit: json.tip_commit,
             indexed_at: json.indexed_at,
             remote: json.remote,
+            repo_path: json.repo_path,
         }
     }
 }
@@ -60,11 +66,64 @@ impl From<IndexedBranchInfoJson> for IndexedBranchInfo {
 /// Store for managing indexed branch records
 pub struct IndexedBranchStore {
     connection: Connection,
+    repo_identity: String,
+    is_local_db: bool,
 }
 
 impl IndexedBranchStore {
-    pub fn new(connection: Connection) -> Self {
-        Self { connection }
+    pub fn new(connection: Connection, repo_identity: String, is_local_db: bool) -> Self {
+        Self {
+            connection,
+            repo_identity,
+            is_local_db,
+        }
+    }
+
+    /// Construct SQL filter string for LanceDB queries based on repository identity and database locality.
+    /// Returns "true" (noop) if the table schema on disk does not have a repo_path column (legacy schema).
+    pub fn build_sql_filter(&self, schema: &Schema) -> String {
+        if schema.column_with_name("repo_path").is_none() {
+            return "true".to_string();
+        }
+        let escaped_rp = Self::escape_sql_literal(&self.repo_identity);
+        if self.is_local_db {
+            format!("(repo_path = '{escaped_rp}' OR repo_path IS NULL)")
+        } else {
+            format!("repo_path = '{escaped_rp}'")
+        }
+    }
+
+    async fn open_table_with_schema(&self) -> Result<(lancedb::Table, Arc<Schema>)> {
+        let table = self
+            .connection
+            .open_table("indexed_branches")
+            .execute()
+            .await?;
+        let schema = table.schema().await?;
+        Ok((table, schema))
+    }
+
+    async fn add_repo_path_column(&self, table: &lancedb::Table) -> Result<()> {
+        tracing::info!("Adding missing 'repo_path' column to indexed_branches table during write");
+        table
+            .add_columns(
+                lancedb::table::NewColumnTransform::SqlExpressions(vec![(
+                    "repo_path".into(),
+                    "CAST(NULL AS string)".into(),
+                )]),
+                None,
+            )
+            .await?;
+
+        let _ = table
+            .create_index(
+                &["repo_path"],
+                LanceIndex::BTree(BTreeIndexBuilder::default()),
+            )
+            .execute()
+            .await;
+
+        Ok(())
     }
 
     /// Get the Arrow schema for the indexed_branches table
@@ -74,25 +133,32 @@ impl IndexedBranchStore {
             Field::new("tip_commit", DataType::Utf8, false),
             Field::new("indexed_at", DataType::Int64, false),
             Field::new("remote", DataType::Utf8, true),
+            Field::new("repo_path", DataType::Utf8, true),
         ]))
     }
 
+    fn escape_sql_literal(val: &str) -> String {
+        val.replace("'", "''")
+    }
+
     /// Record that a branch has been indexed at a specific commit
     pub async fn record_branch_indexed(&self, info: &IndexedBranchInfo) -> Result<()> {
-        // First, remove any existing record for this branch
+        // First, remove any existing record for this branch under current repo scoping
         self.remove_branch(&info.branch_name).await?;
 
-        let table = self
-            .connection
-            .open_table("indexed_branches")
-            .execute()
-            .await?;
+        let (table, schema) = self.open_table_with_schema().await?;
+
+        // Migrate the schema to support scoping if writing to an older version
+        if schema.column_with_name("repo_path").is_none() {
+            self.add_repo_path_column(&table).await?;
+        }
 
         // Build arrays for each column
         let mut branch_name_builder = StringBuilder::new();
         let mut tip_commit_builder = StringBuilder::new();
         let mut indexed_at_builder = arrow::array::Int64Builder::new();
         let mut remote_builder = StringBuilder::new();
+        let mut repo_path_builder = StringBuilder::new();
 
         branch_name_builder.append_value(&info.branch_name);
         tip_commit_builder.append_value(&info.tip_commit);
@@ -101,6 +167,10 @@ impl IndexedBranchStore {
             Some(r) => remote_builder.append_value(r),
             None => remote_builder.append_null(),
         }
+        match &info.repo_path {
+            Some(rp) => repo_path_builder.append_value(rp),
+            None => repo_path_builder.append_null(),
+        }
 
         let batch = RecordBatch::try_from_iter(vec![
             (
@@ -116,6 +186,10 @@ impl IndexedBranchStore {
                 Arc::new(indexed_at_builder.finish()) as ArrayRef,
             ),
             ("remote", Arc::new(remote_builder.finish()) as ArrayRef),
+            (
+                "repo_path",
+                Arc::new(repo_path_builder.finish()) as ArrayRef,
+            ),
         ])?;
 
         table.add(vec![batch]).execute().await?;
@@ -131,45 +205,46 @@ impl IndexedBranchStore {
 
     /// Get full information about a specific branch
     pub async fn get_branch_info(&self, branch_name: &str) -> Result<Option<IndexedBranchInfo>> {
-        let table = self
-            .connection
-            .open_table("indexed_branches")
-            .execute()
-            .await?;
+        let (table, schema) = self.open_table_with_schema().await?;
 
-        let escaped_name = branch_name.replace("'", "''");
-        let filter = format!("branch_name = '{escaped_name}'");
+        let escaped_name = Self::escape_sql_literal(branch_name);
+        let repo_filter = self.build_sql_filter(&schema);
+        let sql_filter = format!("{repo_filter} AND branch_name = '{escaped_name}'");
 
         let results = table
             .query()
-            .only_if(filter)
-            .limit(1)
+            .only_if(sql_filter)
             .execute()
             .await?
             .try_collect::<Vec<_>>()
             .await?;
 
-        if results.is_empty() || results[0].num_rows() == 0 {
+        if results.is_empty() {
             return Ok(None);
         }
 
-        self.extract_record_from_batch(&results[0], 0)
+        // Disambiguate between legacy un-scoped (repo_path = NULL) and repo-scoped records
+        // by selecting the record with the most recent indexed_at timestamp.
+        let newest = results
+            .iter()
+            .flat_map(|batch| (0..batch.num_rows()).map(move |i| (batch, i)))
+            .map(|(batch, i)| self.extract_record_from_batch(batch, i))
+            .collect::<Result<Vec<_>, _>>()?
+            .into_iter()
+            .flatten()
+            .max_by_key(|info| info.indexed_at);
+
+        Ok(newest)
     }
 
     /// List all indexed branches
     pub async fn list_indexed_branches(&self) -> Result<Vec<IndexedBranchInfo>> {
-        let table = self
-            .connection
-            .open_table("indexed_branches")
-            .execute()
-            .await?;
+        let (table, schema) = self.open_table_with_schema().await?;
 
-        let results = table
-            .query()
-            .execute()
-            .await?
-            .try_collect::<Vec<_>>()
-            .await?;
+        let sql_filter = self.build_sql_filter(&schema);
+        let query = table.query().only_if(sql_filter);
+
+        let results = query.execute().await?.try_collect::<Vec<_>>().await?;
 
         let mut branches = Vec::new();
         for batch in &results {
@@ -197,16 +272,13 @@ impl IndexedBranchStore {
 
     /// Remove a branch record (used when branch is deleted or before updating)
     pub async fn remove_branch(&self, branch_name: &str) -> Result<()> {
-        let table = self
-            .connection
-            .open_table("indexed_branches")
-            .execute()
-            .await?;
+        let (table, schema) = self.open_table_with_schema().await?;
 
-        let escaped_name = branch_name.replace("'", "''");
-        let filter = format!("branch_name = '{escaped_name}'");
+        let escaped_name = Self::escape_sql_literal(branch_name);
+        let repo_filter = self.build_sql_filter(&schema);
+        let sql_filter = format!("{repo_filter} AND branch_name = '{escaped_name}'");
 
-        table.delete(&filter).await?;
+        table.delete(&sql_filter).await?;
         Ok(())
     }
 
@@ -227,18 +299,15 @@ impl IndexedBranchStore {
 
     /// Get all branches that point to a specific commit
     pub async fn get_branches_at_commit(&self, commit_sha: &str) -> Result<Vec<IndexedBranchInfo>> {
-        let table = self
-            .connection
-            .open_table("indexed_branches")
-            .execute()
-            .await?;
+        let (table, schema) = self.open_table_with_schema().await?;
 
-        let escaped_sha = commit_sha.replace("'", "''");
-        let filter = format!("tip_commit = '{escaped_sha}'");
+        let escaped_sha = Self::escape_sql_literal(commit_sha);
+        let repo_filter = self.build_sql_filter(&schema);
+        let sql_filter = format!("{repo_filter} AND tip_commit = '{escaped_sha}'");
 
         let results = table
             .query()
-            .only_if(filter)
+            .only_if(sql_filter)
             .execute()
             .await?
             .try_collect::<Vec<_>>()
@@ -256,18 +325,14 @@ impl IndexedBranchStore {
         Ok(branches)
     }
 
-    /// Get total count of indexed branches
+    /// Get total count of indexed branches matching repository filter
     pub async fn count(&self) -> Result<usize> {
-        let table = self
-            .connection
-            .open_table("indexed_branches")
-            .execute()
-            .await?;
-        Ok(table.count_rows(None).await?)
+        let branches = self.list_indexed_branches().await?;
+        Ok(branches.len())
     }
 
     /// Extract an IndexedBranchInfo from a batch at the given row index
-    fn extract_record_from_batch(
+    pub fn extract_record_from_batch(
         &self,
         batch: &RecordBatch,
         row: usize,
@@ -300,6 +365,13 @@ impl IndexedBranchStore {
             .downcast_ref::<arrow::array::StringArray>()
             .ok_or_else(|| anyhow::anyhow!("Invalid remote column type"))?;
 
+        // Extract optional repo_path if column exists and value at row is non-null
+        let repo_path = batch
+            .column_by_name("repo_path")
+            .and_then(|col| col.as_any().downcast_ref::<arrow::array::StringArray>())
+            .filter(|arr| !arr.is_null(row))
+            .map(|arr| arr.value(row).to_string());
+
         let branch_name = branch_name_array.value(row).to_string();
         let tip_commit = tip_commit_array.value(row).to_string();
         let indexed_at = indexed_at_array.value(row);
@@ -314,6 +386,7 @@ impl IndexedBranchStore {
             tip_commit,
             indexed_at,
             remote,
+            repo_path,
         }))
     }
 }
@@ -337,7 +410,7 @@ mod tests {
             .await
             .unwrap();
 
-        let store = IndexedBranchStore::new(connection);
+        let store = IndexedBranchStore::new(connection, "~/projects/repoA/.git".to_string(), true);
         (tmpdir, store)
     }
 
@@ -350,6 +423,7 @@ mod tests {
             tip_commit: "abc123def456".to_string(),
             indexed_at: 1699900000,
             remote: None,
+            repo_path: None,
         };
 
         store.record_branch_indexed(&info).await.unwrap();
@@ -372,6 +446,7 @@ mod tests {
             tip_commit: "789abc123".to_string(),
             indexed_at: 1699900100,
             remote: Some("origin".to_string()),
+            repo_path: None,
         };
 
         store.record_branch_indexed(&info).await.unwrap();
@@ -392,6 +467,7 @@ mod tests {
             tip_commit: "commit1".to_string(),
             indexed_at: 1699900000,
             remote: None,
+            repo_path: None,
         };
         store.record_branch_indexed(&info1).await.unwrap();
 
@@ -401,6 +477,7 @@ mod tests {
             tip_commit: "commit2".to_string(),
             indexed_at: 1699900100,
             remote: None,
+            repo_path: None,
         };
         store.record_branch_indexed(&info2).await.unwrap();
 
@@ -423,18 +500,21 @@ mod tests {
                 tip_commit: "commit1".to_string(),
                 indexed_at: 1699900000,
                 remote: None,
+                repo_path: None,
             },
             IndexedBranchInfo {
                 branch_name: "develop".to_string(),
                 tip_commit: "commit2".to_string(),
                 indexed_at: 1699900100,
                 remote: None,
+                repo_path: None,
             },
             IndexedBranchInfo {
                 branch_name: "origin/feature".to_string(),
                 tip_commit: "commit3".to_string(),
                 indexed_at: 1699900200,
                 remote: Some("origin".to_string()),
+                repo_path: None,
             },
         ];
 
@@ -460,6 +540,7 @@ mod tests {
             tip_commit: "abc123".to_string(),
             indexed_at: 1699900000,
             remote: None,
+            repo_path: None,
         };
         store.record_branch_indexed(&info).await.unwrap();
 
@@ -480,6 +561,7 @@ mod tests {
             tip_commit: "abc123".to_string(),
             indexed_at: 1699900000,
             remote: None,
+            repo_path: None,
         };
         store.record_branch_indexed(&info).await.unwrap();
 
@@ -500,24 +582,28 @@ mod tests {
                 tip_commit: "commit1".to_string(),
                 indexed_at: 1699900000,
                 remote: None,
+                repo_path: None,
             },
             IndexedBranchInfo {
                 branch_name: "origin/main".to_string(),
                 tip_commit: "commit2".to_string(),
                 indexed_at: 1699900100,
                 remote: Some("origin".to_string()),
+                repo_path: None,
             },
             IndexedBranchInfo {
                 branch_name: "origin/develop".to_string(),
                 tip_commit: "commit3".to_string(),
                 indexed_at: 1699900200,
                 remote: Some("origin".to_string()),
+                repo_path: None,
             },
             IndexedBranchInfo {
                 branch_name: "upstream/main".to_string(),
                 tip_commit: "commit4".to_string(),
                 indexed_at: 1699900300,
                 remote: Some("upstream".to_string()),
+                repo_path: None,
             },
         ];
 
@@ -545,18 +631,21 @@ mod tests {
                 tip_commit: shared_commit.to_string(),
                 indexed_at: 1699900000,
                 remote: None,
+                repo_path: None,
             },
             IndexedBranchInfo {
                 branch_name: "release".to_string(),
                 tip_commit: shared_commit.to_string(),
                 indexed_at: 1699900100,
                 remote: None,
+                repo_path: None,
             },
             IndexedBranchInfo {
                 branch_name: "develop".to_string(),
                 tip_commit: "different456".to_string(),
                 indexed_at: 1699900200,
                 remote: None,
+                repo_path: None,
             },
         ];
 
@@ -579,6 +668,7 @@ mod tests {
             tip_commit: "abc123".to_string(),
             indexed_at: 1699900000,
             remote: None,
+            repo_path: None,
         };
         store.record_branch_indexed(&info).await.unwrap();
 
@@ -599,6 +689,7 @@ mod tests {
             tip_commit: "abc123".to_string(),
             indexed_at: 1699900000,
             remote: None,
+            repo_path: None,
         };
         store.record_branch_indexed(&info).await.unwrap();
 
@@ -609,4 +700,153 @@ mod tests {
         assert!(retrieved.is_some());
         assert_eq!(retrieved.unwrap().branch_name, "feature/user's-branch");
     }
+
+    #[tokio::test]
+    async fn test_shared_db_branch_isolation() {
+        let tmpdir = tempfile::tempdir().unwrap();
+        let connection = lancedb::connect(tmpdir.path().to_str().unwrap())
+            .execute()
+            .await
+            .unwrap();
+
+        let schema = IndexedBranchStore::get_schema();
+        let empty_batch = RecordBatch::new_empty(schema.clone());
+        connection
+            .create_table("indexed_branches", vec![empty_batch])
+            .execute()
+            .await
+            .unwrap();
+
+        let repo_a = "~/projects/repoA/.git";
+        let repo_b = "~/projects/repoB/.git";
+
+        let store_a = IndexedBranchStore::new(connection.clone(), repo_a.to_string(), false);
+        let store_b = IndexedBranchStore::new(connection.clone(), repo_b.to_string(), false);
+
+        let info_a = IndexedBranchInfo {
+            branch_name: "main".to_string(),
+            tip_commit: "commit_a_123".to_string(),
+            indexed_at: 1699900000,
+            remote: None,
+            repo_path: Some(repo_a.to_string()),
+        };
+        store_a.record_branch_indexed(&info_a).await.unwrap();
+
+        let info_b = IndexedBranchInfo {
+            branch_name: "main".to_string(),
+            tip_commit: "commit_b_456".to_string(),
+            indexed_at: 1699900100,
+            remote: None,
+            repo_path: Some(repo_b.to_string()),
+        };
+        store_b.record_branch_indexed(&info_b).await.unwrap();
+
+        // In shared mode, store A sees only repo A's main branch
+        let branch_a = store_a.get_branch_info("main").await.unwrap().unwrap();
+        assert_eq!(branch_a.tip_commit, "commit_a_123");
+
+        // Store B sees only repo B's main branch
+        let branch_b = store_b.get_branch_info("main").await.unwrap().unwrap();
+        assert_eq!(branch_b.tip_commit, "commit_b_456");
+
+        // Listing for Store A yields 1 branch
+        let list_a = store_a.list_indexed_branches().await.unwrap();
+        assert_eq!(list_a.len(), 1);
+        assert_eq!(list_a[0].tip_commit, "commit_a_123");
+    }
+
+    #[tokio::test]
+    async fn test_local_db_matches_null_and_active() {
+        let tmpdir = tempfile::tempdir().unwrap();
+        let connection = lancedb::connect(tmpdir.path().to_str().unwrap())
+            .execute()
+            .await
+            .unwrap();
+
+        let schema = IndexedBranchStore::get_schema();
+        let empty_batch = RecordBatch::new_empty(schema.clone());
+        connection
+            .create_table("indexed_branches", vec![empty_batch])
+            .execute()
+            .await
+            .unwrap();
+
+        let repo_a = "~/projects/repoA/.git";
+
+        let store_local = IndexedBranchStore::new(connection.clone(), repo_a.to_string(), true);
+        let store_shared = IndexedBranchStore::new(connection.clone(), repo_a.to_string(), false);
+
+        // Record a legacy NULL branch entry
+        let info_null = IndexedBranchInfo {
+            branch_name: "legacy_branch".to_string(),
+            tip_commit: "legacy_sha".to_string(),
+            indexed_at: 1699900000,
+            remote: None,
+            repo_path: None,
+        };
+        store_local.record_branch_indexed(&info_null).await.unwrap();
+
+        // In Local mode, the legacy NULL branch is returned
+        let branch = store_local
+            .get_branch_info("legacy_branch")
+            .await
+            .unwrap()
+            .unwrap();
+        assert_eq!(branch.tip_commit, "legacy_sha");
+
+        // In Shared mode, the legacy NULL branch is NOT returned
+        let branch_shared = store_shared.get_branch_info("legacy_branch").await.unwrap();
+        assert!(branch_shared.is_none());
+    }
+
+    #[tokio::test]
+    async fn test_read_old_indexed_branches_table_does_not_alter_schema() {
+        let dir = tempfile::tempdir().unwrap();
+        let db_path = dir.path().to_str().unwrap();
+
+        // 1. Manually create an old 4-column indexed_branches table FIRST
+        let connection = lancedb::connect(db_path).execute().await.unwrap();
+        let old_schema = Arc::new(Schema::new(vec![
+            Field::new("branch_name", DataType::Utf8, false),
+            Field::new("tip_commit", DataType::Utf8, false),
+            Field::new("indexed_at", DataType::Int64, false),
+            Field::new("remote", DataType::Utf8, true),
+        ]));
+
+        let empty_batch = RecordBatch::new_empty(old_schema);
+        connection
+            .create_table("indexed_branches", vec![empty_batch])
+            .execute()
+            .await
+            .unwrap();
+
+        // 2. NOW open IndexedBranchStore on the old schema
+        let store = IndexedBranchStore::new(connection, "~/projects/repoA/.git".to_string(), true);
+
+        // Reading branches works cleanly without altering disk schema
+        let (_table, schema_read) = store.open_table_with_schema().await.unwrap();
+        assert!(
+            schema_read.column_with_name("repo_path").is_none(),
+            "Read-only operations should not alter schema on disk"
+        );
+
+        let branches = store.list_indexed_branches().await.unwrap();
+        assert!(branches.is_empty());
+
+        // Writing a new branch record triggers schema migration on demand
+        let info = IndexedBranchInfo {
+            branch_name: "main".to_string(),
+            tip_commit: "sha123".to_string(),
+            indexed_at: 1699900000,
+            remote: None,
+            repo_path: None,
+        };
+        store.record_branch_indexed(&info).await.unwrap();
+
+        let (_table, schema_after_write) = store.open_table_with_schema().await.unwrap();
+        assert!(
+            schema_after_write.column_with_name("repo_path").is_some(),
+            "Write operation should add repo_path column on demand"
+        );
+    }
 }
diff --git a/src/database/connection.rs b/src/database/connection.rs
index 0542e0f..5477bbe 100644
--- a/src/database/connection.rs
+++ b/src/database/connection.rs
@@ -24,13 +24,22 @@ use crate::types::{FunctionInfo, TypeInfo, TypedefInfo};
 use crate::vectorizer::CodeVectorizer;
 use crate::workdir::WorkdirIndex;
 use std::collections::HashSet;
+use std::path::PathBuf;
 
 // Optimal batch size for LanceDB operations
 pub const OPTIMAL_BATCH_SIZE: usize = 65536;
 
+#[derive(Debug, Default, Clone, PartialEq, Eq)]
+pub struct GcStats {
+    pub deleted_stale_branch_records: usize,
+}
+
 pub struct DatabaseManager {
     connection: Connection,
-    git_repo_path: String,
+    db_path: String,
+    work_dir: String,
+    repo_identity: String,
+    is_local_db: bool,
     function_store: FunctionStore,
     type_store: TypeStore,
     typedef_store: TypedefStore,
@@ -45,22 +54,26 @@ pub struct DatabaseManager {
 }
 
 impl DatabaseManager {
-    pub async fn new(db_path: &str, git_repo_path: String) -> Result<Self> {
+    pub async fn new(db_path: &str, work_dir: String) -> Result<Self> {
         let connection = lancedb::connect(db_path).execute().await?;
+        let (repo_identity, is_local_db) = crate::git::determine_db_locality(&work_dir, db_path);
 
         Ok(Self {
             connection: connection.clone(),
-            git_repo_path: git_repo_path.clone(),
+            db_path: db_path.to_string(),
+            work_dir: work_dir.clone(),
+            repo_identity: repo_identity.clone(),
+            is_local_db,
             function_store: FunctionStore::new(connection.clone()),
             type_store: TypeStore::new(connection.clone()),
             typedef_store: TypedefStore::new(connection.clone()),
-            search_manager: SearchManager::new(connection.clone(), git_repo_path),
+            search_manager: SearchManager::new(connection.clone(), work_dir),
             vector_search_manager: VectorSearchManager::new(connection.clone()),
             schema_manager: SchemaManager::new(connection.clone()),
             processed_file_store: ProcessedFileStore::new(connection.clone()),
             content_store: ContentStore::new(connection.clone()),
             symbol_filename_store: SymbolFilenameStore::new(connection.clone()),
-            branch_store: IndexedBranchStore::new(connection.clone()),
+            branch_store: IndexedBranchStore::new(connection.clone(), repo_identity, is_local_db),
             workdir_index: std::sync::RwLock::new(None),
         })
     }
@@ -504,16 +517,30 @@ impl DatabaseManager {
         self.schema_manager.compact_and_cleanup().await
     }
 
+    pub async fn compact_lore_tables(&self) -> Result<()> {
+        self.schema_manager.compact_lore_tables().await
+    }
+
     /// Drop and recreate all tables for maximum space savings
     pub async fn drop_and_recreate_tables(&self) -> Result<()> {
         self.schema_manager.drop_and_recreate_tables().await
     }
 
-    /// Create FTS indices for lore table (must be called after data is inserted)
+    /// Drop and rebuild all FTS indices for the lore table from scratch.
     pub async fn create_lore_fts_indices(&self) -> Result<()> {
         self.schema_manager.create_lore_fts_indices().await
     }
 
+    /// Create FTS indices only if they do not already exist.
+    pub async fn ensure_lore_fts_indices(&self) -> Result<()> {
+        self.schema_manager.ensure_lore_fts_indices().await
+    }
+
+    /// Merge newly-inserted rows into existing lore FTS indices.
+    pub async fn optimize_lore_fts_indices(&self) -> Result<()> {
+        self.schema_manager.optimize_lore_fts_indices().await
+    }
+
     /// Get lore table information including row count and indices
     pub async fn get_lore_table_info(&self) -> Result<String> {
         use std::fmt::Write;
@@ -2592,14 +2619,14 @@ impl DatabaseManager {
         file_paths: &[String],
         git_sha: &str,
     ) -> Result<std::collections::HashMap<String, String>> {
-        match crate::git::resolve_files_at_commit(&self.git_repo_path, git_sha, file_paths) {
+        match crate::git::resolve_files_at_commit(&self.work_dir, git_sha, file_paths) {
             Ok(resolved_hashes) => {
                 // If no files were resolved, log this as a warning
                 if resolved_hashes.is_empty() {
                     tracing::warn!(
                         "No files were resolved at commit {} in repository {}",
                         git_sha,
-                        self.git_repo_path
+                        self.work_dir
                     );
                 }
 
@@ -2607,7 +2634,7 @@ impl DatabaseManager {
             }
             Err(e) => {
                 tracing::error!("DatabaseManager::resolve_git_file_hashes: Failed to resolve git files at commit {}: {}", git_sha, e);
-                tracing::error!("Repository path: {}", self.git_repo_path);
+                tracing::error!("Repository path: {}", self.work_dir);
                 tracing::error!("Requested file paths: {:?}", file_paths);
                 Ok(std::collections::HashMap::new()) // Return empty map instead of failing, let caller handle
             }
@@ -2699,6 +2726,14 @@ impl DatabaseManager {
 
     // ==================== Branch Management ====================
 
+    pub fn repo_identity(&self) -> &str {
+        &self.repo_identity
+    }
+
+    pub fn is_local_db(&self) -> bool {
+        self.is_local_db
+    }
+
     /// Record that a branch has been indexed at a specific commit
     pub async fn record_branch_indexed(
         &self,
@@ -2707,6 +2742,7 @@ impl DatabaseManager {
         remote: Option<&str>,
     ) -> Result<()> {
         use crate::database::branches::IndexedBranchInfo;
+        let repo_path = (!self.is_local_db).then(|| self.repo_identity.clone());
         let info = IndexedBranchInfo {
             branch_name: branch_name.to_string(),
             tip_commit: tip_commit.to_string(),
@@ -2715,6 +2751,7 @@ impl DatabaseManager {
                 .unwrap()
                 .as_secs() as i64,
             remote: remote.map(|s| s.to_string()),
+            repo_path,
         };
         self.branch_store.record_branch_indexed(&info).await
     }
@@ -2769,6 +2806,93 @@ impl DatabaseManager {
         self.branch_store.count().await
     }
 
+    /// Performs garbage collection on stale branch records belonging to deleted repository paths.
+    pub async fn garbage_collect(&self) -> Result<GcStats> {
+        let mut stats = GcStats::default();
+
+        let table = match self
+            .connection
+            .open_table("indexed_branches")
+            .execute()
+            .await
+        {
+            Ok(t) => t,
+            Err(_) => return Ok(stats),
+        };
+
+        use futures::TryStreamExt;
+        let results = table
+            .query()
+            .only_if("repo_path IS NOT NULL")
+            .execute()
+            .await?
+            .try_collect::<Vec<_>>()
+            .await?;
+
+        let mut known_repo_paths = HashSet::new();
+        for batch in &results {
+            if let Some(col) = batch.column_by_name("repo_path") {
+                if let Some(arr) = col.as_any().downcast_ref::<arrow::array::StringArray>() {
+                    for i in 0..batch.num_rows() {
+                        if !arr.is_null(i) {
+                            known_repo_paths.insert(arr.value(i).to_string());
+                        }
+                    }
+                }
+            }
+        }
+
+        let mut repo_paths_to_clean = Vec::new();
+        for rp in known_repo_paths {
+            let expanded_path = if let Some(stripped) = rp.strip_prefix("~/") {
+                if let Some(home) = dirs::home_dir() {
+                    home.join(stripped)
+                } else {
+                    PathBuf::from(&rp)
+                }
+            } else {
+                PathBuf::from(&rp)
+            };
+
+            if !expanded_path.exists() {
+                repo_paths_to_clean.push(rp);
+            }
+        }
+
+        for stale_rp in repo_paths_to_clean {
+            let escaped = stale_rp.replace("'", "''");
+            let filter = format!("repo_path = '{escaped}'");
+            let _ = table.delete(&filter).await;
+            stats.deleted_stale_branch_records += 1;
+        }
+
+        // If the database directory does not reside within any Git repository,
+        // it is a standalone/central database that can never be local to any repository.
+        // In this case, un-scoped legacy (repo_path IS NULL) records can never be read
+        // and are safe to clean up.
+        let db_dir = std::path::Path::new(&self.db_path)
+            .parent()
+            .unwrap_or_else(|| std::path::Path::new("."));
+        let is_inside_any_git_repo = gix::discover(db_dir).is_ok();
+
+        if !is_inside_any_git_repo {
+            let null_results = table
+                .query()
+                .only_if("repo_path IS NULL")
+                .execute()
+                .await?
+                .try_collect::<Vec<_>>()
+                .await?;
+            let null_count: usize = null_results.iter().map(|b| b.num_rows()).sum();
+            if null_count > 0 {
+                let _ = table.delete("repo_path IS NULL").await;
+                stats.deleted_stale_branch_records += null_count;
+            }
+        }
+
+        Ok(stats)
+    }
+
     // ==================== End Branch Management ====================
 
     pub async fn get_existing_function_names(&self) -> Result<std::collections::HashSet<String>> {
@@ -2826,7 +2950,7 @@ impl DatabaseManager {
         );
 
         // Step 1: Get directly changed files between commits
-        let changed_files = get_changed_files(&self.git_repo_path, commit_a, commit_b)?;
+        let changed_files = get_changed_files(&self.work_dir, commit_a, commit_b)?;
         let mut files_to_scan = std::collections::HashSet::new();
 
         // Add all directly changed files (except deleted ones)
@@ -3288,7 +3412,7 @@ impl DatabaseManager {
             }
             None => {
                 // Default to current commit
-                match crate::git::get_git_sha(&self.git_repo_path) {
+                match crate::git::get_git_sha(&self.work_dir) {
                     Ok(Some(current_sha)) => {
                         tracing::info!("Using current git commit SHA: {}", current_sha);
                         current_sha
@@ -3423,16 +3547,12 @@ impl DatabaseManager {
         let mut manifest = std::collections::HashMap::new();
 
         // Use shared tree traversal utility
-        crate::git::walk_tree_at_commit(
-            &self.git_repo_path,
-            git_sha,
-            |relative_path, object_id| {
-                // Normalize path by removing any double slashes
-                let normalized_path = relative_path.replace("//", "/");
-                manifest.insert(normalized_path, object_id.to_string());
-                Ok(())
-            },
-        )?;
+        crate::git::walk_tree_at_commit(&self.work_dir, git_sha, |relative_path, object_id| {
+            // Normalize path by removing any double slashes
+            let normalized_path = relative_path.replace("//", "/");
+            manifest.insert(normalized_path, object_id.to_string());
+            Ok(())
+        })?;
 
         // Merge with workdir overlay (adds dirty files, removes deleted files)
         Ok(self.workdir_merged_manifest(manifest))
@@ -3933,13 +4053,19 @@ impl DatabaseManager {
         Ok(())
     }
 
-    /// Insert lore emails into the database
-    pub async fn insert_lore_emails(&self, emails: &[crate::types::LoreEmailInfo]) -> Result<()> {
+    /// Insert lore emails into the database, returning the indices of
+    /// any emails that could not be stored.  Callers must not record
+    /// commit SHAs for failed indices in the lore_indexed_commits
+    /// table, otherwise those emails are permanently lost.
+    pub async fn insert_lore_emails(
+        &self,
+        emails: &[crate::types::LoreEmailInfo],
+    ) -> Result<Vec<usize>> {
         use arrow::datatypes::{DataType, Field, Schema};
         use std::sync::Arc;
 
         if emails.is_empty() {
-            return Ok(());
+            return Ok(Vec::new());
         }
 
         tracing::info!(
@@ -3961,10 +4087,10 @@ impl DatabaseManager {
             Field::new("symbols", DataType::Utf8, false),
         ]));
 
-        // Deduplicate by message_id within the batch. A lore archive
-        // can contain the same email in multiple git commits, and
-        // LanceDB merge_insert requires each target row to be matched
-        // by at most one source row. Keep the last occurrence.
+        // Deduplicate by message_id within the batch.  A lore
+        // archive can contain the same email in multiple git commits;
+        // appending duplicates would create redundant rows.  Keep the
+        // last occurrence.
         let mut seen = std::collections::HashMap::with_capacity(emails.len());
         for (i, email) in emails.iter().enumerate() {
             seen.insert(&email.message_id, i);
@@ -3983,38 +4109,50 @@ impl DatabaseManager {
 
         let table = self.connection.open_table("lore").execute().await?;
 
-        // Try inserting the full batch first -- a single merge_insert
-        // is far cheaper than many small ones because each call is a
-        // full read-modify-write cycle in LanceDB.  Fall back to
-        // chunked insertion only when the batch exhausts DataFusion's
-        // memory pool (the RepartitionExec OOM described in the
-        // comment below).
-        //
-        // Retrying the full set of indices on failure is safe:
-        // merge_insert is not transactional, so a failed call may
-        // have persisted some rows before the error.  Because the
-        // upsert key is message_id, re-inserting those rows is
-        // idempotent.
-        if let Err(e) = Self::merge_insert_lore_chunk(&table, emails, &dedup_indices, &schema).await
-        {
+        let mut failed_indices: Vec<usize> = Vec::new();
+
+        // Filter out emails whose message_id already exists in the
+        // table.  Normally every email here is new because
+        // index_lore_archive skips already-indexed commits, but a
+        // partial failure on a previous run can leave rows in the
+        // table whose commit SHA was never recorded in
+        // lore_indexed_commits.  Filtering avoids duplicates without
+        // relying on merge_insert, which hits a spurious null-column
+        // error in lance-file 3.0.1.
+        let new_indices = Self::filter_existing_lore_ids(&table, emails, &dedup_indices).await?;
+
+        if new_indices.len() < dedup_indices.len() {
+            tracing::info!(
+                "insert_lore_emails: {} of {} already in table, inserting {}",
+                dedup_indices.len() - new_indices.len(),
+                dedup_indices.len(),
+                new_indices.len(),
+            );
+        }
+
+        if new_indices.is_empty() {
+            tracing::info!("insert_lore_emails: no new emails to insert");
+            return Ok(failed_indices);
+        }
+
+        // Use table.add() instead of merge_insert.  merge_insert in
+        // lance-file 3.0.1 introduces spurious nulls during its
+        // internal join/write phase, causing every insert to fail
+        // with "subject contained null values".  Plain append avoids
+        // the merge codepath entirely.  Duplicates are prevented by
+        // the filter_existing_lore_ids check above.
+        if let Err(e) = Self::add_lore_chunk(&table, emails, &new_indices, &schema).await {
             tracing::warn!(
                 "insert_lore_emails: full batch of {} failed ({}), \
                  falling back to chunked insertion",
-                dedup_indices.len(),
+                new_indices.len(),
                 e
             );
 
-            // Lore emails carry full bodies, so each row is large
-            // compared to code-analysis records.  LanceDB
-            // merge_insert uses DataFusion's RepartitionExec, whose
-            // memory pool can be exhausted by a single oversized
-            // RecordBatch.  Insert in sub-batches to bound peak
-            // memory per operation.
             const MAX_CHUNK: usize = 128;
 
-            for chunk in dedup_indices.chunks(MAX_CHUNK) {
-                if let Err(e) = Self::merge_insert_lore_chunk(&table, emails, chunk, &schema).await
-                {
+            for chunk in new_indices.chunks(MAX_CHUNK) {
+                if let Err(e) = Self::add_lore_chunk(&table, emails, chunk, &schema).await {
                     tracing::warn!(
                         "insert_lore_emails: chunk of {} failed ({}), \
                          retrying individually",
@@ -4022,8 +4160,7 @@ impl DatabaseManager {
                         e
                     );
                     for &idx in chunk {
-                        if let Err(e2) =
-                            Self::merge_insert_lore_chunk(&table, emails, &[idx], &schema).await
+                        if let Err(e2) = Self::add_lore_chunk(&table, emails, &[idx], &schema).await
                         {
                             tracing::warn!(
                                 "insert_lore_emails: skipping \
@@ -4031,20 +4168,80 @@ impl DatabaseManager {
                                 emails[idx].message_id,
                                 e2
                             );
+                            failed_indices.push(idx);
                         }
                     }
                 }
             }
         }
 
+        if !failed_indices.is_empty() {
+            tracing::warn!(
+                "insert_lore_emails: {} of {} emails failed to insert",
+                failed_indices.len(),
+                new_indices.len()
+            );
+        }
+
         tracing::info!("insert_lore_emails: Batch insertion complete");
 
-        Ok(())
+        Ok(failed_indices)
+    }
+
+    /// Return the subset of `indices` whose message_id does not
+    /// already exist in the lore table.
+    async fn filter_existing_lore_ids(
+        table: &lancedb::Table,
+        emails: &[crate::types::LoreEmailInfo],
+        indices: &[usize],
+    ) -> Result<Vec<usize>> {
+        use futures::TryStreamExt as _;
+
+        // Build an IN-list predicate.  message_ids are already
+        // validated to be non-empty, but escape single quotes for
+        // the SQL literal.
+        let id_list: Vec<String> = indices
+            .iter()
+            .map(|&i| format!("'{}'", emails[i].message_id.replace('\'', "''")))
+            .collect();
+
+        let predicate = format!("message_id IN ({})", id_list.join(", "));
+
+        let stream = table
+            .query()
+            .select(lancedb::query::Select::Columns(vec![
+                "message_id".to_string()
+            ]))
+            .only_if(&predicate)
+            .execute()
+            .await?;
+        let batches: Vec<_> = stream.try_collect().await?;
+
+        let mut existing = std::collections::HashSet::new();
+        for batch in &batches {
+            if let Some(col) = batch.column_by_name("message_id") {
+                let arr = col
+                    .as_any()
+                    .downcast_ref::<arrow::array::StringArray>()
+                    .expect("message_id column must be StringArray");
+                for i in 0..arr.len() {
+                    if !arr.is_null(i) {
+                        existing.insert(arr.value(i).to_string());
+                    }
+                }
+            }
+        }
+
+        Ok(indices
+            .iter()
+            .copied()
+            .filter(|&i| !existing.contains(&emails[i].message_id))
+            .collect())
     }
 
     /// Build a [`RecordBatch`] from the given email indices and
-    /// merge-insert it into the lore table.
-    async fn merge_insert_lore_chunk(
+    /// append it to the lore table.
+    async fn add_lore_chunk(
         table: &lancedb::Table,
         emails: &[crate::types::LoreEmailInfo],
         indices: &[usize],
@@ -4098,15 +4295,7 @@ impl DatabaseManager {
         ];
 
         let batch = RecordBatch::try_new(schema.clone(), columns)?;
-        let batches = vec![Ok(batch)];
-        let batch_iterator =
-            arrow::record_batch::RecordBatchIterator::new(batches.into_iter(), schema.clone());
-
-        let mut merge_insert = table.merge_insert(&["message_id"]);
-        merge_insert
-            .when_matched_update_all(None)
-            .when_not_matched_insert_all();
-        merge_insert.execute(Box::new(batch_iterator)).await?;
+        table.add(vec![batch]).execute().await?;
         Ok(())
     }
 
@@ -4293,18 +4482,42 @@ impl DatabaseManager {
         let mut first_iteration = true;
 
         loop {
-            let fts_query =
-                FullTextSearchQuery::new(fts_pattern.clone()).with_column(field.to_owned())?;
+            // When the FTS pattern is empty (e.g. regex ".*" has no
+            // alphanumeric tokens), skip FTS and use a plain table
+            // scan so the regex post-filter still runs.
+            let batches: Vec<_> = if fts_pattern.is_empty() {
+                tracing::info!(
+                    "FTS pattern empty for field '{}', falling back to table scan",
+                    field
+                );
+                let mut query_builder = table.query();
+                if let Some(ref filter) = date_filter {
+                    query_builder = query_builder.only_if(filter);
+                }
+                query_builder
+                    .limit(fts_limit)
+                    .execute()
+                    .await?
+                    .try_collect()
+                    .await?
+            } else {
+                let fts_query =
+                    FullTextSearchQuery::new(fts_pattern.clone()).with_column(field.to_owned())?;
 
-            let mut query_builder = table.query().full_text_search(fts_query);
+                let mut query_builder = table.query().full_text_search(fts_query);
 
-            // Apply date filter at database level so limit applies to date-filtered results
-            if let Some(ref filter) = date_filter {
-                query_builder = query_builder.only_if(filter);
-            }
+                // Apply date filter at database level so limit applies to date-filtered results
+                if let Some(ref filter) = date_filter {
+                    query_builder = query_builder.only_if(filter);
+                }
 
-            let stream = query_builder.limit(fts_limit).execute().await?;
-            let batches: Vec<_> = stream.try_collect().await?;
+                query_builder
+                    .limit(fts_limit)
+                    .execute()
+                    .await?
+                    .try_collect()
+                    .await?
+            };
 
             let fts_count: usize = batches.iter().map(|b| b.num_rows()).sum();
             tracing::info!(
@@ -4468,6 +4681,7 @@ impl DatabaseManager {
 
     /// Helper function to query lore emails by multiple fields and return intersection
     /// Uses regex alternation for OR within fields, intersection for AND across fields
+    #[allow(clippy::too_many_arguments)]
     pub(crate) async fn query_lore_by_fields_intersection(
         &self,
         from_patterns: Option<&[String]>,
@@ -4475,18 +4689,35 @@ impl DatabaseManager {
         body_patterns: Option<&[String]>,
         recipients_patterns: Option<&[String]>,
         search_limit: usize,
+        since_date: Option<&str>,
+        until_date: Option<&str>,
     ) -> Result<std::collections::HashSet<String>> {
         use std::collections::HashSet;
 
         let lore_table = self.connection.open_table("lore").execute().await?;
         let mut field_result_sets: Vec<HashSet<String>> = Vec::new();
 
-        // Helper function to query a field using FTS with regex post-filtering
+        // Parse date filters into DateTime for temporal comparison
+        // in query_field_impl (RFC 2822 string comparison is not
+        // meaningful for date ordering).
+        let since_dt = since_date
+            .and_then(|d| chrono::DateTime::parse_from_rfc2822(d).ok())
+            .map(|dt| dt.with_timezone(&chrono::Utc));
+        let until_dt = until_date
+            .and_then(|d| chrono::DateTime::parse_from_rfc2822(d).ok())
+            .map(|dt| dt.with_timezone(&chrono::Utc));
+
+        // Helper function to query a field using FTS with regex and
+        // date post-filtering.  Selects the "date" column alongside
+        // the searched field so temporal filtering happens on the
+        // already-fetched FTS candidates without extra lookups.
         async fn query_field_impl(
             lore_table: &lancedb::Table,
             field_name: String,
             pattern: String,
             search_limit: usize,
+            since: Option<chrono::DateTime<chrono::Utc>>,
+            until: Option<chrono::DateTime<chrono::Utc>>,
         ) -> Result<HashSet<String>> {
             // FTS uses simple tokenizer - normalize pattern by stripping special chars
             let fts_pattern = pattern
@@ -4502,28 +4733,52 @@ impl DatabaseManager {
                 fts_pattern
             );
 
-            let fts_query =
-                FullTextSearchQuery::new(fts_pattern).with_column(field_name.clone())?;
-            let mut query = lore_table.query().full_text_search(fts_query).select(
-                lancedb::query::Select::Columns(vec![
-                    "message_id".to_string(),
-                    "_score".to_string(),
-                    field_name.clone(),
-                ]),
-            );
-
-            // Apply limit - use large limit if search_limit is 0 (unlimited)
-            // FTS has a default limit of 10, so we must explicitly set a large limit
             let effective_limit = if search_limit > 0 {
                 search_limit
             } else {
                 100000
             };
-            query = query.limit(effective_limit);
 
-            let results = query.execute().await?.try_collect::<Vec<_>>().await?;
+            // When the FTS pattern is empty (e.g. regex ".*" has no
+            // alphanumeric tokens), skip FTS and fall back to a plain
+            // table scan so the regex post-filter still runs.
+            let results = if fts_pattern.is_empty() {
+                tracing::info!(
+                    "FTS pattern empty for field '{}', falling back to table scan",
+                    field_name
+                );
+                lore_table
+                    .query()
+                    .select(lancedb::query::Select::Columns(vec![
+                        "message_id".to_string(),
+                        field_name.clone(),
+                        "date".to_string(),
+                    ]))
+                    .limit(effective_limit)
+                    .execute()
+                    .await?
+                    .try_collect::<Vec<_>>()
+                    .await?
+            } else {
+                let fts_query =
+                    FullTextSearchQuery::new(fts_pattern).with_column(field_name.clone())?;
+                lore_table
+                    .query()
+                    .full_text_search(fts_query)
+                    .select(lancedb::query::Select::Columns(vec![
+                        "message_id".to_string(),
+                        "_score".to_string(),
+                        field_name.clone(),
+                        "date".to_string(),
+                    ]))
+                    .limit(effective_limit)
+                    .execute()
+                    .await?
+                    .try_collect::<Vec<_>>()
+                    .await?
+            };
 
-            // Step 2: Post-filter with regex in memory
+            // Post-filter with regex and date range in memory
             let fts_result_count: usize = results.iter().map(|b| b.num_rows()).sum();
             tracing::info!(
                 "FTS returned {} candidates for field '{}'",
@@ -4534,37 +4789,46 @@ impl DatabaseManager {
             let regex = regex::RegexBuilder::new(&pattern)
                 .case_insensitive(true)
                 .build()?;
+            let has_date_filter = since.is_some() || until.is_some();
             let mut message_ids = HashSet::new();
-            let mut first_candidate_logged = false;
+            let mut bad_dates: usize = 0;
 
             for batch in &results {
-                let msg_array = batch
-                    .column(0)
-                    .as_any()
-                    .downcast_ref::<arrow::array::StringArray>()
-                    .unwrap();
-                let field_array = batch
-                    .column(2)
-                    .as_any()
-                    .downcast_ref::<arrow::array::StringArray>()
-                    .unwrap();
+                let msg_array: &arrow::array::StringArray = super::get_column(batch, "message_id")?;
+                let field_array: &arrow::array::StringArray =
+                    super::get_column(batch, &field_name)?;
+                let date_array: &arrow::array::StringArray = super::get_column(batch, "date")?;
 
                 for i in 0..batch.num_rows() {
-                    let field_value = field_array.value(i);
-
-                    // Log first candidate for debugging
-                    if !first_candidate_logged {
-                        tracing::info!("Sample {} value: '{}'", field_name, field_value);
-                        tracing::info!("Regex pattern: '{}'", pattern);
-                        first_candidate_logged = true;
+                    if !regex.is_match(field_array.value(i)) {
+                        continue;
                     }
-
-                    if regex.is_match(field_value) {
-                        message_ids.insert(msg_array.value(i).to_string());
+                    if has_date_filter {
+                        if let Ok(dt) = chrono::DateTime::parse_from_rfc2822(date_array.value(i)) {
+                            let dt_utc = dt.with_timezone(&chrono::Utc);
+                            if since.is_some_and(|s| dt_utc < s) {
+                                continue;
+                            }
+                            if until.is_some_and(|u| dt_utc > u) {
+                                continue;
+                            }
+                        } else {
+                            bad_dates += 1;
+                            continue;
+                        }
                     }
+                    message_ids.insert(msg_array.value(i).to_string());
                 }
             }
 
+            if bad_dates > 0 {
+                tracing::warn!(
+                    "Skipped {} candidates with unparseable dates for field '{}'",
+                    bad_dates,
+                    field_name
+                );
+            }
+
             tracing::info!(
                 "Regex filter kept {} of {} FTS candidates",
                 message_ids.len(),
@@ -4574,71 +4838,90 @@ impl DatabaseManager {
             Ok(message_ids)
         }
 
-        // Query from field
+        // Query from field (OR across patterns, then push single set)
         if let Some(patterns) = from_patterns {
             if !patterns.is_empty() {
-                // FTS: join patterns with spaces for multi-keyword search
-                let combined_pattern = patterns.join(" ");
-                let results = query_field_impl(
-                    &lore_table,
-                    "from".to_string(),
-                    combined_pattern,
-                    search_limit,
-                )
-                .await?;
-                tracing::info!("lore from field returned {} results", results.len());
-                field_result_sets.push(results);
+                let mut field_union = HashSet::new();
+                for pattern in patterns {
+                    let results = query_field_impl(
+                        &lore_table,
+                        "from".to_string(),
+                        pattern.clone(),
+                        search_limit,
+                        since_dt,
+                        until_dt,
+                    )
+                    .await?;
+                    field_union.extend(results);
+                }
+                tracing::info!("lore from field returned {} results", field_union.len());
+                field_result_sets.push(field_union);
             }
         }
 
-        // Query subject field
+        // Query subject field (OR across patterns)
         if let Some(patterns) = subject_patterns {
             if !patterns.is_empty() {
-                // FTS: join patterns with spaces for multi-keyword search
-                let combined_pattern = patterns.join(" ");
-                let results = query_field_impl(
-                    &lore_table,
-                    "subject".to_string(),
-                    combined_pattern,
-                    search_limit,
-                )
-                .await?;
-                tracing::info!("lore subject field returned {} results", results.len());
-                field_result_sets.push(results);
+                let mut field_union = HashSet::new();
+                for pattern in patterns {
+                    let results = query_field_impl(
+                        &lore_table,
+                        "subject".to_string(),
+                        pattern.clone(),
+                        search_limit,
+                        since_dt,
+                        until_dt,
+                    )
+                    .await?;
+                    field_union.extend(results);
+                }
+                tracing::info!("lore subject field returned {} results", field_union.len());
+                field_result_sets.push(field_union);
             }
         }
 
-        // Query body field
+        // Query body field (OR across patterns)
         if let Some(patterns) = body_patterns {
             if !patterns.is_empty() {
-                // FTS: join patterns with spaces for multi-keyword search
-                let combined_pattern = patterns.join(" ");
-                let results = query_field_impl(
-                    &lore_table,
-                    "body".to_string(),
-                    combined_pattern,
-                    search_limit,
-                )
-                .await?;
-                tracing::info!("lore body field returned {} results", results.len());
-                field_result_sets.push(results);
+                let mut field_union = HashSet::new();
+                for pattern in patterns {
+                    let results = query_field_impl(
+                        &lore_table,
+                        "body".to_string(),
+                        pattern.clone(),
+                        search_limit,
+                        since_dt,
+                        until_dt,
+                    )
+                    .await?;
+                    field_union.extend(results);
+                }
+                tracing::info!("lore body field returned {} results", field_union.len());
+                field_result_sets.push(field_union);
             }
         }
 
-        // Query recipients field
+        // Query recipients field (OR across patterns)
         if let Some(patterns) = recipients_patterns {
             if !patterns.is_empty() {
-                // FTS: join patterns with spaces for multi-keyword search
-                let combined_pattern = patterns.join(" ");
-                let results = query_field_impl(
-                    &lore_table,
-                    "recipients".to_string(),
-                    combined_pattern,
-                    search_limit,
-                )
-                .await?;
-                tracing::info!("lore recipients field returned {} results", results.len());
-                field_result_sets.push(results);
+                let mut field_union = HashSet::new();
+                for pattern in patterns {
+                    let results = query_field_impl(
+                        &lore_table,
+                        "recipients".to_string(),
+                        pattern.clone(),
+                        search_limit,
+                        since_dt,
+                        until_dt,
+                    )
+                    .await?;
+                    field_union.extend(results);
+                }
+                tracing::info!(
+                    "lore recipients field returned {} results",
+                    field_union.len()
+                );
+                field_result_sets.push(field_union);
             }
         }
 
@@ -4704,7 +4987,9 @@ impl DatabaseManager {
         let body_patterns = field_map.get("body").map(|v| v.as_slice());
         let recipients_patterns = field_map.get("recipients").map(|v| v.as_slice());
 
-        // Use helper to get intersection of message_ids
+        // Use helper to get intersection of message_ids.
+        // Date range is pushed into FTS queries so the candidate set
+        // is already bounded before intersection and fetching.
         let intersection = self
             .query_lore_by_fields_intersection(
                 from_patterns,
@@ -4712,6 +4997,8 @@ impl DatabaseManager {
                 body_patterns,
                 recipients_patterns,
                 0, // No limit for individual queries
+                since_date,
+                until_date,
             )
             .await?;
 
@@ -4721,54 +5008,15 @@ impl DatabaseManager {
             return Ok(Vec::new());
         }
 
-        // Parse filter dates to Unix timestamps for database-level filtering
-        let since_timestamp = since_date
-            .and_then(|d| chrono::DateTime::parse_from_rfc2822(d).ok())
-            .map(|dt| dt.timestamp());
-        let until_timestamp = until_date
-            .and_then(|d| chrono::DateTime::parse_from_rfc2822(d).ok())
-            .map(|dt| dt.timestamp());
+        // Fetch full email records for the intersection.
+        // Date filtering was already pushed into the per-field FTS
+        // queries, so the intersection is already date-bounded.
+        let count_limit = if limit > 0 { limit } else { intersection.len() };
+        let ids: Vec<&String> = intersection.iter().take(count_limit).collect();
+        let final_emails = self.fetch_lore_emails_by_message_ids(&ids).await?;
 
         tracing::info!(
-            "lore multi_field search: since_timestamp={:?} until_timestamp={:?}",
-            since_timestamp,
-            until_timestamp
-        );
-
-        // Build date filter clause — only if the column exists in the table
-        let has_date_timestamp = {
-            let table = self.connection.open_table("lore").execute().await?;
-            table
-                .schema()
-                .await
-                .map(|s| s.field_with_name("date_timestamp").is_ok())
-                .unwrap_or(false)
-        };
-        let date_filter = if has_date_timestamp {
-            match (since_timestamp, until_timestamp) {
-                (Some(since), Some(until)) => Some(format!(
-                    "date_timestamp >= {} AND date_timestamp <= {}",
-                    since, until
-                )),
-                (Some(since), None) => Some(format!("date_timestamp >= {}", since)),
-                (None, Some(until)) => Some(format!("date_timestamp <= {}", until)),
-                (None, None) => None,
-            }
-        } else {
-            None
-        };
-
-        // Fetch emails with date filtering at database level
-        let final_emails = self
-            .fetch_lore_emails_by_message_ids_with_filter(
-                &intersection,
-                date_filter.as_deref(),
-                limit,
-            )
-            .await?;
-
-        tracing::info!(
-            "Fetched {} email records matching date filter from {} message_ids",
+            "Fetched {} email records from {} message_ids",
             final_emails.len(),
             intersection.len()
         );
@@ -4776,13 +5024,13 @@ impl DatabaseManager {
         Ok(final_emails)
     }
 
-    /// Fetch lore emails by message IDs with optional SQL filter and limit
-    /// This allows efficient database-level filtering (e.g., by date_timestamp)
-    async fn fetch_lore_emails_by_message_ids_with_filter(
+    /// Fetch lore emails by message IDs in batches.
+    ///
+    /// Builds `message_id IN (...)` predicates in chunks to avoid
+    /// per-ID round-trips while keeping predicate size bounded.
+    async fn fetch_lore_emails_by_message_ids(
         &self,
-        message_ids: &std::collections::HashSet<String>,
-        filter: Option<&str>,
-        limit: usize,
+        message_ids: &[&String],
     ) -> Result<Vec<crate::types::LoreEmailInfo>> {
         use arrow::array::AsArray;
         use futures::TryStreamExt;
@@ -4792,88 +5040,99 @@ impl DatabaseManager {
         }
 
         let table = self.connection.open_table("lore").execute().await?;
+        let mut emails = Vec::with_capacity(message_ids.len());
 
-        // Build message_id IN clause
-        let escaped_ids: Vec<String> = message_ids
-            .iter()
-            .map(|id| {
-                let escaped = id.replace('\'', "''");
-                format!("'{}'", escaped)
-            })
-            .collect();
-        let in_clause = format!("message_id IN ({})", escaped_ids.join(", "));
-
-        // Combine with optional filter
-        let where_clause = match filter {
-            Some(f) => format!("{} AND {}", in_clause, f),
-            None => in_clause,
-        };
-
-        tracing::debug!("Lore query with filter: {}", where_clause);
+        for chunk in message_ids.chunks(500) {
+            let placeholders: Vec<String> = chunk
+                .iter()
+                .map(|id| {
+                    let escaped = id.replace('\'', "''");
+                    format!("'{}'", escaped)
+                })
+                .collect();
+            let predicate = format!("message_id IN ({})", placeholders.join(", "));
 
-        // Query with filter, ordered by date_timestamp descending (newest first)
-        let effective_limit = if limit > 0 { limit } else { message_ids.len() };
-        let results = table
-            .query()
-            .only_if(&where_clause)
-            .limit(effective_limit)
-            .execute()
-            .await?
-            .try_collect::<Vec<_>>()
-            .await?;
+            let results = table
+                .query()
+                .only_if(&predicate)
+                .limit(chunk.len())
+                .execute()
+                .await?
+                .try_collect::<Vec<_>>()
+                .await?;
 
-        let mut emails = Vec::new();
-        for batch in results {
-            let git_commit_sha_col = batch.column_by_name("git_commit_sha").unwrap();
-            let from_col = batch.column_by_name("from").unwrap();
-            let date_col = batch.column_by_name("date").unwrap();
-            let date_timestamp_col = batch.column_by_name("date_timestamp");
-            let message_id_col = batch.column_by_name("message_id").unwrap();
-            let in_reply_to_col = batch.column_by_name("in_reply_to").unwrap();
-            let subject_col = batch.column_by_name("subject").unwrap();
-            let references_col = batch.column_by_name("references").unwrap();
-            let recipients_col = batch.column_by_name("recipients").unwrap();
-            let body_col = batch.column_by_name("body").unwrap();
-            let symbols_col = batch.column_by_name("symbols").unwrap();
-
-            let git_commit_sha_arr = git_commit_sha_col.as_string::<i32>();
-            let from_arr = from_col.as_string::<i32>();
-            let date_arr = date_col.as_string::<i32>();
-            let date_timestamp_arr = date_timestamp_col
-                .and_then(|c| c.as_any().downcast_ref::<arrow::array::Int64Array>());
-            let message_id_arr = message_id_col.as_string::<i32>();
-            let in_reply_to_arr = in_reply_to_col.as_string::<i32>();
-            let subject_arr = subject_col.as_string::<i32>();
-            let references_arr = references_col.as_string::<i32>();
-            let recipients_arr = recipients_col.as_string::<i32>();
-            let body_arr = body_col.as_string::<i32>();
-            let symbols_arr = symbols_col.as_string::<i32>();
+            for batch in &results {
+                let git_commit_shas = batch
+                    .column_by_name("git_commit_sha")
+                    .ok_or_else(|| anyhow::anyhow!("Missing git_commit_sha column"))?
+                    .as_string::<i32>();
+                let from_addrs = batch
+                    .column_by_name("from")
+                    .ok_or_else(|| anyhow::anyhow!("Missing from column"))?
+                    .as_string::<i32>();
+                let dates = batch
+                    .column_by_name("date")
+                    .ok_or_else(|| anyhow::anyhow!("Missing date column"))?
+                    .as_string::<i32>();
+                let date_timestamps = batch
+                    .column_by_name("date_timestamp")
+                    .and_then(|c| c.as_any().downcast_ref::<arrow::array::Int64Array>());
+                let msg_ids = batch
+                    .column_by_name("message_id")
+                    .ok_or_else(|| anyhow::anyhow!("Missing message_id column"))?
+                    .as_string::<i32>();
+                let in_reply_tos = batch
+                    .column_by_name("in_reply_to")
+                    .ok_or_else(|| anyhow::anyhow!("Missing in_reply_to column"))?
+                    .as_string::<i32>();
+                let subjects = batch
+                    .column_by_name("subject")
+                    .ok_or_else(|| anyhow::anyhow!("Missing subject column"))?
+                    .as_string::<i32>();
+                let references_list = batch
+                    .column_by_name("references")
+                    .ok_or_else(|| anyhow::anyhow!("Missing references column"))?
+                    .as_string::<i32>();
+                let recipients_list = batch
+                    .column_by_name("recipients")
+                    .ok_or_else(|| anyhow::anyhow!("Missing recipients column"))?
+                    .as_string::<i32>();
+                let bodies = batch
+                    .column_by_name("body")
+                    .ok_or_else(|| anyhow::anyhow!("Missing body column"))?
+                    .as_string::<i32>();
+                let symbols_list = batch
+                    .column_by_name("symbols")
+                    .ok_or_else(|| anyhow::anyhow!("Missing symbols column"))?
+                    .as_string::<i32>();
 
-            for i in 0..batch.num_rows() {
-                let symbols_json = symbols_arr.value(i);
-                let symbols: Vec<String> = serde_json::from_str(symbols_json).unwrap_or_default();
+                for i in 0..batch.num_rows() {
+                    let symbols_json = symbols_list.value(i);
+                    let symbols: Vec<String> =
+                        serde_json::from_str(symbols_json).unwrap_or_default();
 
-                emails.push(crate::types::LoreEmailInfo {
-                    git_commit_sha: git_commit_sha_arr.value(i).to_string(),
-                    from: from_arr.value(i).to_string(),
-                    date: date_arr.value(i).to_string(),
-                    date_timestamp: Self::get_date_timestamp(date_timestamp_arr, date_arr, i),
-                    message_id: message_id_arr.value(i).to_string(),
-                    in_reply_to: if in_reply_to_arr.is_null(i) {
-                        None
-                    } else {
-                        Some(in_reply_to_arr.value(i).to_string())
-                    },
-                    subject: subject_arr.value(i).to_string(),
-                    references: if references_arr.is_null(i) {
-                        None
-                    } else {
-                        Some(references_arr.value(i).to_string())
-                    },
-                    recipients: recipients_arr.value(i).to_string(),
-                    body: body_arr.value(i).to_string(),
-                    symbols,
-                });
+                    emails.push(crate::types::LoreEmailInfo {
+                        git_commit_sha: git_commit_shas.value(i).to_string(),
+                        from: from_addrs.value(i).to_string(),
+                        date: dates.value(i).to_string(),
+                        date_timestamp: Self::get_date_timestamp(date_timestamps, dates, i),
+                        message_id: msg_ids.value(i).to_string(),
+                        in_reply_to: if in_reply_tos.is_null(i) {
+                            None
+                        } else {
+                            Some(in_reply_tos.value(i).to_string())
+                        },
+                        subject: subjects.value(i).to_string(),
+                        references: if references_list.is_null(i) {
+                            None
+                        } else {
+                            Some(references_list.value(i).to_string())
+                        },
+                        recipients: recipients_list.value(i).to_string(),
+                        body: bodies.value(i).to_string(),
+                        symbols,
+                    });
+                }
             }
         }
 
@@ -5781,3 +6040,53 @@ impl DatabaseManager {
         Ok(())
     }
 }
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+    use crate::database::branches::{IndexedBranchInfo, IndexedBranchStore};
+    use arrow::array::RecordBatch;
+
+    #[tokio::test]
+    async fn test_garbage_collect_null_records_in_standalone_db() {
+        let tmpdir = tempfile::tempdir().unwrap();
+        let db_path = tmpdir.path().join("standalone.db");
+
+        let git_dir = tempfile::tempdir().unwrap();
+        let _ = std::process::Command::new("git")
+            .args(["init", git_dir.path().to_str().unwrap()])
+            .output();
+
+        let db = DatabaseManager::new(
+            db_path.to_str().unwrap(),
+            git_dir.path().to_str().unwrap().to_string(),
+        )
+        .await
+        .unwrap();
+
+        let schema = IndexedBranchStore::get_schema();
+        let empty_batch = RecordBatch::new_empty(schema);
+        db.connection()
+            .create_table("indexed_branches", vec![empty_batch])
+            .execute()
+            .await
+            .unwrap();
+
+        let store_local = IndexedBranchStore::new(
+            db.connection().clone(),
+            git_dir.path().to_str().unwrap().to_string(),
+            true,
+        );
+        let info_null = IndexedBranchInfo {
+            branch_name: "legacy".to_string(),
+            tip_commit: "sha1".to_string(),
+            indexed_at: 100,
+            remote: None,
+            repo_path: None,
+        };
+        store_local.record_branch_indexed(&info_null).await.unwrap();
+
+        let stats = db.garbage_collect().await.unwrap();
+        assert_eq!(stats.deleted_stale_branch_records, 1);
+    }
+}
diff --git a/src/database/schema.rs b/src/database/schema.rs
index 28977e2..b4a4026 100644
--- a/src/database/schema.rs
+++ b/src/database/schema.rs
@@ -1,5 +1,6 @@
 // SPDX-License-Identifier: MIT OR Apache-2.0
 use anyhow::Result;
+use arrow::array::Array;
 use arrow::datatypes::{DataType, Field, Schema};
 use arrow::record_batch::RecordBatch;
 use futures::{stream, StreamExt, TryStreamExt};
@@ -14,7 +15,7 @@ use std::sync::Arc;
 pub enum OptimizeOutcome {
     /// Table was successfully optimized (all operations completed)
     Optimized,
-    /// Table was skipped (too small to benefit from optimization)
+    /// Table was skipped (e.g., too few rows to benefit)
     Skipped,
     /// Optimization was attempted but one or more operations failed
     PartialFailure,
@@ -301,7 +302,7 @@ impl SchemaManager {
                     tracing::info!("Lore table migration complete");
                 }
                 OptimizeOutcome::Skipped => {
-                    tracing::info!("Lore table compaction skipped (table too small)");
+                    tracing::info!("Lore table compaction skipped (preserving FTS indices)");
                 }
                 OptimizeOutcome::PartialFailure => {
                     tracing::warn!("Lore table compaction partially failed");
@@ -309,6 +310,109 @@ impl SchemaManager {
             }
         }
 
+        // Add the date_timestamp column if missing. Databases created
+        // before this column was introduced have a 10-column schema;
+        // merge_insert of 11-column batches silently fails, causing
+        // new emails to be skipped while their commit SHAs are still
+        // recorded as indexed.
+        if schema.column_with_name("date_timestamp").is_none() {
+            tracing::info!("Migrating lore table: adding 'date_timestamp' column");
+            table
+                .add_columns(
+                    lancedb::table::NewColumnTransform::SqlExpressions(vec![(
+                        "date_timestamp".into(),
+                        "CAST(0 AS BIGINT)".into(),
+                    )]),
+                    None,
+                )
+                .await?;
+
+            // Purge lore_indexed_commits so that previously-skipped
+            // emails are re-examined on the next --lore refresh.
+            self.reconcile_lore_indexed_commits().await?;
+        }
+
+        Ok(())
+    }
+
+    /// Remove entries from lore_indexed_commits whose git_commit_sha
+    /// does not appear in the lore table.  This recovers from the
+    /// schema-mismatch bug where SHAs were recorded as indexed but the
+    /// corresponding emails were never stored.
+    async fn reconcile_lore_indexed_commits(&self) -> Result<()> {
+        let lore = self.connection.open_table("lore").execute().await?;
+        let idx = self
+            .connection
+            .open_table("lore_indexed_commits")
+            .execute()
+            .await?;
+
+        // Collect the set of SHAs actually present in the lore table.
+        let lore_stream = lore
+            .query()
+            .select(lancedb::query::Select::Columns(vec![
+                "git_commit_sha".to_string()
+            ]))
+            .execute()
+            .await?;
+        let lore_batches: Vec<_> = lore_stream.try_collect().await?;
+
+        let mut lore_shas = std::collections::HashSet::new();
+        for batch in &lore_batches {
+            if let Some(col) = batch.column_by_name("git_commit_sha") {
+                if let Some(arr) = col.as_any().downcast_ref::<arrow::array::StringArray>() {
+                    for i in 0..arr.len() {
+                        lore_shas.insert(arr.value(i).to_string());
+                    }
+                }
+            }
+        }
+
+        // Collect SHAs from lore_indexed_commits.
+        let idx_stream = idx
+            .query()
+            .select(lancedb::query::Select::Columns(vec![
+                "git_commit_sha".to_string()
+            ]))
+            .execute()
+            .await?;
+        let idx_batches: Vec<_> = idx_stream.try_collect().await?;
+
+        let mut orphaned: Vec<String> = Vec::new();
+        for batch in &idx_batches {
+            if let Some(col) = batch.column_by_name("git_commit_sha") {
+                if let Some(arr) = col.as_any().downcast_ref::<arrow::array::StringArray>() {
+                    for i in 0..arr.len() {
+                        let sha = arr.value(i);
+                        if !lore_shas.contains(sha) {
+                            orphaned.push(sha.to_string());
+                        }
+                    }
+                }
+            }
+        }
+
+        if orphaned.is_empty() {
+            tracing::info!("reconcile_lore_indexed_commits: no orphaned entries");
+            return Ok(());
+        }
+
+        tracing::info!(
+            "reconcile_lore_indexed_commits: removing {} orphaned entries",
+            orphaned.len()
+        );
+
+        // Delete in chunks to avoid oversized SQL predicates.
+        for chunk in orphaned.chunks(500) {
+            let placeholders: Vec<String> = chunk
+                .iter()
+                .map(|s| format!("'{}'", s.replace('\'', "''")))
+                .collect();
+            let predicate = format!("git_commit_sha IN ({})", placeholders.join(", "));
+            idx.delete(&predicate).await?;
+        }
+
+        tracing::info!("reconcile_lore_indexed_commits: done");
         Ok(())
     }
 
@@ -788,38 +892,11 @@ impl SchemaManager {
         }
     }
 
-    async fn try_create_fts_index(
-        &self,
-        table: &lancedb::table::Table,
-        columns: &[&str],
-        description: &str,
-    ) {
-        // Configure FTS index optimized for source code and technical docs
-        let fts_config = FtsIndexBuilder::default()
-            .with_position(false) // Enable phrase queries for exact code snippets
-            .base_tokenizer("simple".to_string())
-            .lower_case(true) // Case-insensitive search for better matching
-            .stem(false) // No stemming (run != running in code)
-            .remove_stop_words(false) // Keep all tokens (if, for, while are keywords)
-            .ascii_folding(true) // Preserve exact characters
-            .max_token_length(Some(100)); // Allow long identifiers/symbols
-
-        tracing::info!("Starting {}", description);
-        match table
-            .create_index(columns, LanceIndex::FTS(fts_config))
-            .execute()
-            .await
-        {
-            Ok(_) => tracing::info!("✓ Completed {}", description),
-            Err(e) => tracing::warn!("Failed to create {}: {}", description, e),
-        }
-    }
-
-    /// Create FTS indices for lore table (must be called after data is inserted)
+    /// Drop and rebuild all FTS indices for the lore table from scratch.
     ///
-    /// Drops any existing FTS indices first, creates fresh ones, then
-    /// prunes so that the on-disk index directories from prior data
-    /// layouts are actually removed.
+    /// Intended for schema migrations and --clear rebuilds where the
+    /// table structure has changed.  Normal incremental indexing should
+    /// use ensure_lore_fts_indices() + optimize_lore_fts_indices().
     pub async fn create_lore_fts_indices(&self) -> Result<()> {
         let table = self.connection.open_table("lore").execute().await?;
 
@@ -841,29 +918,9 @@ impl SchemaManager {
             }
         }
 
-        // Create FTS indices for text search on all searchable fields in parallel
-        let start_time = std::time::Instant::now();
-        tracing::info!("Creating 5 FTS indices for lore table in parallel (from, subject, body, recipients, symbols)...");
-
-        // Create all 5 indices concurrently for faster indexing
-        tokio::join!(
-            self.try_create_fts_index(&table, &["from"], "FTS index on lore.from"),
-            self.try_create_fts_index(&table, &["subject"], "FTS index on lore.subject"),
-            self.try_create_fts_index(&table, &["body"], "FTS index on lore.body"),
-            self.try_create_fts_index(&table, &["recipients"], "FTS index on lore.recipients"),
-            self.try_create_fts_index(&table, &["symbols"], "FTS index on lore.symbols"),
-        );
+        Self::create_all_fts_indices(&table).await?;
 
-        let elapsed = start_time.elapsed();
-        tracing::info!(
-            "Completed creating 5 FTS indices in {:.1}s",
-            elapsed.as_secs_f64()
-        );
-
-        // Prune orphaned index data left behind by drop_index()
-        // and by OptimizeAction::Index in optimize_single_table(),
-        // which rebuilds all indices (including FTS) into new
-        // directories without removing the old ones.
+        // Prune orphaned index data left behind by drop_index().
         if dropped {
             tracing::info!("Pruning orphaned index data from lore table...");
             if let Err(e) = table
@@ -883,6 +940,149 @@ impl SchemaManager {
         Ok(())
     }
 
+    /// Create FTS indices only if they do not already exist.
+    ///
+    /// After the first full build, subsequent indexing runs call this
+    /// to ensure the indices are present, then optimize_lore_fts_indices()
+    /// to merge newly-inserted rows into the existing indices.
+    pub async fn ensure_lore_fts_indices(&self) -> Result<()> {
+        use lancedb::index::IndexType;
+        let table = self.connection.open_table("lore").execute().await?;
+        let indices: Vec<lancedb::index::IndexConfig> =
+            (table.list_indices().await).unwrap_or_default();
+
+        let fts_count = indices
+            .iter()
+            .filter(|idx| idx.index_type == IndexType::FTS)
+            .count();
+
+        // All 5 FTS indices present — nothing to do.
+        if fts_count >= 5 {
+            tracing::info!(
+                "Lore FTS indices already present ({} indices), skipping creation",
+                fts_count
+            );
+            return Ok(());
+        }
+
+        if fts_count > 0 {
+            tracing::info!(
+                "Only {} of 5 FTS indices present, rebuilding all",
+                fts_count
+            );
+            // Drop the partial set so create_index does not collide.
+            for idx in &indices {
+                if idx.index_type == IndexType::FTS {
+                    let _ = table.drop_index(&idx.name).await;
+                }
+            }
+        } else {
+            tracing::info!("No FTS indices found, creating initial set");
+        }
+
+        Self::create_all_fts_indices(&table).await
+    }
+
+    /// Merge newly-inserted rows into existing lore FTS indices.
+    ///
+    /// LanceDB's native FTS engine serves unindexed rows via a
+    /// brute-force fallback at query time, so queries remain correct
+    /// even before this call.  Running optimize merges those rows
+    /// into the inverted index structure, eliminating the scan cost.
+    pub async fn optimize_lore_fts_indices(&self) -> Result<()> {
+        // Guard against running on a table with a large _indices/
+        // backlog.  lance/index/append.rs opens every delta fragment
+        // for a column before merging any, so peak memory scales
+        // linearly with the number of fragments per column.  On
+        // memory-constrained systems a backlog in the thousands
+        // drives semcode-index into swap and gets it OOM-killed.
+        // Query correctness is preserved regardless: unindexed rows
+        // still fall back to a brute-force scan.
+        const MAX_LORE_INDEX_FRAGMENTS: usize = 100;
+        let uri = self.connection.uri();
+        let indices_dir = std::path::Path::new(uri)
+            .join("lore.lance")
+            .join("_indices");
+        if let Ok(rd) = std::fs::read_dir(&indices_dir) {
+            let count = rd.count();
+            if count > MAX_LORE_INDEX_FRAGMENTS {
+                tracing::warn!(
+                    "Skipping lore FTS index optimization: \
+                     {} _indices/ fragments exceeds {} threshold. \
+                     Queries remain correct via brute-force fallback. \
+                     Rebuild the lore table on a larger host to recover.",
+                    count,
+                    MAX_LORE_INDEX_FRAGMENTS
+                );
+                return Ok(());
+            }
+        }
+
+        let table = self.connection.open_table("lore").execute().await?;
+        let start_time = std::time::Instant::now();
+
+        tracing::info!("Optimizing lore FTS indices (incremental merge)...");
+        table
+            .optimize(OptimizeAction::Index(Default::default()))
+            .await?;
+
+        let elapsed = start_time.elapsed();
+        tracing::info!(
+            "Lore FTS index optimization completed in {:.1}s",
+            elapsed.as_secs_f64()
+        );
+        Ok(())
+    }
+
+    /// Shared helper: create all 5 FTS indices on an already-opened table.
+    async fn create_all_fts_indices(table: &lancedb::table::Table) -> Result<()> {
+        let start_time = std::time::Instant::now();
+        tracing::info!(
+            "Creating 5 FTS indices for lore table (from, subject, body, recipients, symbols)..."
+        );
+
+        tokio::join!(
+            Self::create_one_fts_index(table, &["from"], "FTS index on lore.from"),
+            Self::create_one_fts_index(table, &["subject"], "FTS index on lore.subject"),
+            Self::create_one_fts_index(table, &["body"], "FTS index on lore.body"),
+            Self::create_one_fts_index(table, &["recipients"], "FTS index on lore.recipients"),
+            Self::create_one_fts_index(table, &["symbols"], "FTS index on lore.symbols"),
+        );
+
+        let elapsed = start_time.elapsed();
+        tracing::info!(
+            "Completed creating 5 FTS indices in {:.1}s",
+            elapsed.as_secs_f64()
+        );
+        Ok(())
+    }
+
+    /// Create a single FTS index on the given columns.
+    async fn create_one_fts_index(
+        table: &lancedb::table::Table,
+        columns: &[&str],
+        description: &str,
+    ) {
+        let fts_config = FtsIndexBuilder::default()
+            .with_position(false)
+            .base_tokenizer("simple".to_string())
+            .lower_case(true)
+            .stem(false)
+            .remove_stop_words(false)
+            .ascii_folding(true)
+            .max_token_length(Some(100));
+
+        tracing::info!("Starting {}", description);
+        match table
+            .create_index(columns, LanceIndex::FTS(fts_config))
+            .execute()
+            .await
+        {
+            Ok(_) => tracing::info!("Completed {}", description),
+            Err(e) => tracing::warn!("Failed to create {}: {}", description, e),
+        }
+    }
+
     pub async fn rebuild_indices(&self) -> Result<()> {
         // Rebuild vector index if needed
         let table_names = self.connection.table_names().execute().await?;
@@ -999,7 +1199,7 @@ impl SchemaManager {
         let skipped = tables_skipped.load(Ordering::Relaxed);
         let failed = tables_failed.load(Ordering::Relaxed);
 
-        println!(
+        tracing::info!(
             "    Optimized {} tables{}{}",
             optimized,
             if skipped > 0 {
@@ -1017,6 +1217,43 @@ impl SchemaManager {
         Ok(())
     }
 
+    /// Compact only the tables modified by lore indexing.
+    ///
+    /// The full `compact_and_cleanup` method processes every table in
+    /// the database, including code-index tables and content shards
+    /// that a lore run never touches.  On memory-constrained systems
+    /// the combined working set of those compactions triggers the OOM
+    /// killer.  This method limits work to the two lore tables and
+    /// processes them sequentially to keep peak memory low.
+    pub async fn compact_lore_tables(&self) -> Result<()> {
+        tracing::info!("Running compaction for lore tables...");
+
+        let table_names = self.connection.table_names().execute().await?;
+        let lore_tables = ["lore", "lore_indexed_commits"];
+
+        for name in &lore_tables {
+            if !table_names.iter().any(|n| n == name) {
+                continue;
+            }
+            match Self::optimize_single_table(&self.connection, name).await {
+                Ok(OptimizeOutcome::Optimized) => {
+                    tracing::info!("Compacted table {}", name);
+                }
+                Ok(OptimizeOutcome::Skipped) => {
+                    tracing::info!("Skipped table {} (too small)", name);
+                }
+                Ok(OptimizeOutcome::PartialFailure) => {
+                    tracing::warn!("Partial failure compacting table {}", name);
+                }
+                Err(e) => {
+                    tracing::warn!("Failed to compact table {}: {}", name, e);
+                }
+            }
+        }
+
+        Ok(())
+    }
+
     /// Optimize a single table - runs compact, prune, and index operations
     ///
     /// Tables with fewer than 1000 rows are skipped since the overhead of
@@ -1025,6 +1262,24 @@ impl SchemaManager {
         connection: &Connection,
         table_name: &str,
     ) -> Result<OptimizeOutcome> {
+        // The lore table is indexed incrementally via
+        // ensure_lore_fts_indices() + optimize_lore_fts_indices().
+        // Running the generic optimize path here does no useful work
+        // that those helpers have not already done, and for large
+        // lore archives its Compact phase walks every delta index
+        // fragment under _indices/, holding per-fragment state until
+        // the operation completes.  On memory-constrained systems the
+        // resident set grows into swap and the OOM killer terminates
+        // semcode-index before compaction finishes, leaving fresh
+        // delta fragments behind each time.  Skip the table entirely.
+        if table_name == "lore" {
+            tracing::info!(
+                "Skipping generic optimization for lore table \
+                 (handled by optimize_lore_fts_indices)"
+            );
+            return Ok(OptimizeOutcome::Skipped);
+        }
+
         // Minimum row count for optimization to be worthwhile
         const MIN_ROWS_FOR_OPTIMIZATION: usize = 1000;
 
diff --git a/src/database/search.rs b/src/database/search.rs
index 5528489..7145177 100644
--- a/src/database/search.rs
+++ b/src/database/search.rs
@@ -34,16 +34,16 @@ pub struct FunctionMatch {
 
 pub struct SearchManager {
     connection: Connection,
-    git_repo_path: String,
+    work_dir: String,
     content_store: ContentStore,
 }
 
 impl SearchManager {
-    pub fn new(connection: Connection, git_repo_path: String) -> Self {
+    pub fn new(connection: Connection, work_dir: String) -> Self {
         let content_store = ContentStore::new(connection.clone());
         Self {
             connection,
-            git_repo_path,
+            work_dir,
             content_store,
         }
     }
@@ -62,7 +62,7 @@ impl SearchManager {
             git_sha
         );
 
-        match crate::git::resolve_files_at_commit(&self.git_repo_path, git_sha, file_paths) {
+        match crate::git::resolve_files_at_commit(&self.work_dir, git_sha, file_paths) {
             Ok(resolved_hashes) => {
                 tracing::debug!(
                     "resolve_git_file_hashes: Successfully resolved {} out of {} file paths",
@@ -1787,6 +1787,7 @@ impl VectorSearchManager {
 
     /// Helper to query lore by fields and return intersection of message_ids
     /// Optimized for large result sets with capacity pre-allocation
+    #[allow(clippy::too_many_arguments)]
     async fn query_lore_fields_intersection(
         &self,
         from_patterns: Option<&[String]>,
@@ -1795,18 +1796,35 @@ impl VectorSearchManager {
         symbols_patterns: Option<&[String]>,
         recipients_patterns: Option<&[String]>,
         search_limit: usize,
+        since_date: Option<&str>,
+        until_date: Option<&str>,
     ) -> Result<std::collections::HashSet<String>> {
         use std::collections::HashSet;
 
         let lore_table = self.connection.open_table("lore").execute().await?;
         let mut field_result_sets: Vec<HashSet<String>> = Vec::new();
 
-        // Helper function to query a field using substring matching and collect message_ids efficiently
+        // Parse date filters into DateTime for temporal comparison
+        // in query_field_impl (RFC 2822 string comparison is not
+        // meaningful for date ordering).
+        let since_dt = since_date
+            .and_then(|d| chrono::DateTime::parse_from_rfc2822(d).ok())
+            .map(|dt| dt.with_timezone(&chrono::Utc));
+        let until_dt = until_date
+            .and_then(|d| chrono::DateTime::parse_from_rfc2822(d).ok())
+            .map(|dt| dt.with_timezone(&chrono::Utc));
+
+        // Helper function to query a field using FTS with regex and
+        // date post-filtering.  Selects the "date" column alongside
+        // the searched field so temporal filtering happens on the
+        // already-fetched FTS candidates without extra lookups.
         async fn query_field_impl(
             lore_table: &lancedb::Table,
             field_name: String,
             pattern: String,
             search_limit: usize,
+            since: Option<chrono::DateTime<chrono::Utc>>,
+            until: Option<chrono::DateTime<chrono::Utc>>,
         ) -> Result<HashSet<String>> {
             let start = std::time::Instant::now();
 
@@ -1826,11 +1844,12 @@ impl VectorSearchManager {
 
             let fts_query =
                 FullTextSearchQuery::new(fts_pattern).with_column(field_name.clone())?;
-            let mut query = lore_table.query().full_text_search(fts_query).select(
+            let query = lore_table.query().full_text_search(fts_query).select(
                 lancedb::query::Select::Columns(vec![
                     "message_id".to_string(),
                     "_score".to_string(),
                     field_name.clone(),
+                    "date".to_string(),
                 ]),
             );
 
@@ -1841,38 +1860,56 @@ impl VectorSearchManager {
             } else {
                 100000
             };
-            query = query.limit(effective_limit);
+            let query = query.limit(effective_limit);
 
             let results = query.execute().await?.try_collect::<Vec<_>>().await?;
 
-            // Step 2: Post-filter results with regex in memory (small result set)
+            // Post-filter with regex and date range in memory
             let regex = regex::RegexBuilder::new(&pattern)
                 .case_insensitive(true)
                 .build()?;
+            let has_date_filter = since.is_some() || until.is_some();
             let mut message_ids = HashSet::new();
+            let mut bad_dates: usize = 0;
 
             for batch in &results {
-                let msg_array = batch
-                    .column(0)
-                    .as_any()
-                    .downcast_ref::<arrow::array::StringArray>()
-                    .unwrap();
-                let field_array = batch
-                    .column(2)
-                    .as_any()
-                    .downcast_ref::<arrow::array::StringArray>()
-                    .unwrap();
+                let msg_array: &arrow::array::StringArray = super::get_column(batch, "message_id")?;
+                let field_array: &arrow::array::StringArray =
+                    super::get_column(batch, &field_name)?;
+                let date_array: &arrow::array::StringArray = super::get_column(batch, "date")?;
 
                 for i in 0..batch.num_rows() {
-                    let field_value = field_array.value(i);
-                    if regex.is_match(field_value) {
-                        message_ids.insert(msg_array.value(i).to_string());
+                    if !regex.is_match(field_array.value(i)) {
+                        continue;
                     }
+                    if has_date_filter {
+                        if let Ok(dt) = chrono::DateTime::parse_from_rfc2822(date_array.value(i)) {
+                            let dt_utc = dt.with_timezone(&chrono::Utc);
+                            if since.is_some_and(|s| dt_utc < s) {
+                                continue;
+                            }
+                            if until.is_some_and(|u| dt_utc > u) {
+                                continue;
+                            }
+                        } else {
+                            bad_dates += 1;
+                            continue;
+                        }
+                    }
+                    message_ids.insert(msg_array.value(i).to_string());
                 }
             }
 
+            if bad_dates > 0 {
+                tracing::warn!(
+                    "Skipped {} candidates with unparseable dates for field '{}'",
+                    bad_dates,
+                    field_name
+                );
+            }
+
             tracing::info!(
-                "vlore filter completed: FTS returned {} candidates, regex filtered to {} in {:?}",
+                "vlore filter completed: FTS returned {} candidates, regex+date filtered to {} in {:?}",
                 results.iter().map(|b| b.num_rows()).sum::<usize>(),
                 message_ids.len(),
                 start.elapsed()
@@ -1881,88 +1918,103 @@ impl VectorSearchManager {
             Ok(message_ids)
         }
 
-        // Query from field
+        // Query from field (OR across patterns)
         if let Some(patterns) = from_patterns {
             if !patterns.is_empty() {
-                // FTS: join patterns with spaces for multi-keyword search
-                let combined_pattern = patterns.join(" ");
-                field_result_sets.push(
-                    query_field_impl(
+                let mut field_union = HashSet::new();
+                for pattern in patterns {
+                    let results = query_field_impl(
                         &lore_table,
                         "from".to_string(),
-                        combined_pattern,
+                        pattern.clone(),
                         search_limit,
+                        since_dt,
+                        until_dt,
                     )
-                    .await?,
-                );
+                    .await?;
+                    field_union.extend(results);
+                }
+                field_result_sets.push(field_union);
             }
         }
 
-        // Query subject field
+        // Query subject field (OR across patterns)
         if let Some(patterns) = subject_patterns {
             if !patterns.is_empty() {
-                // FTS: join patterns with spaces for multi-keyword search
-                let combined_pattern = patterns.join(" ");
-                field_result_sets.push(
-                    query_field_impl(
+                let mut field_union = HashSet::new();
+                for pattern in patterns {
+                    let results = query_field_impl(
                         &lore_table,
                         "subject".to_string(),
-                        combined_pattern,
+                        pattern.clone(),
                         search_limit,
+                        since_dt,
+                        until_dt,
                     )
-                    .await?,
-                );
+                    .await?;
+                    field_union.extend(results);
+                }
+                field_result_sets.push(field_union);
             }
         }
 
-        // Query body field
+        // Query body field (OR across patterns)
         if let Some(patterns) = body_patterns {
             if !patterns.is_empty() {
-                // FTS: join patterns with spaces for multi-keyword search
-                let combined_pattern = patterns.join(" ");
-                field_result_sets.push(
-                    query_field_impl(
+                let mut field_union = HashSet::new();
+                for pattern in patterns {
+                    let results = query_field_impl(
                         &lore_table,
                         "body".to_string(),
-                        combined_pattern,
+                        pattern.clone(),
                         search_limit,
+                        since_dt,
+                        until_dt,
                     )
-                    .await?,
-                );
+                    .await?;
+                    field_union.extend(results);
+                }
+                field_result_sets.push(field_union);
             }
         }
 
-        // Query symbols field
+        // Query symbols field (OR across patterns)
         if let Some(patterns) = symbols_patterns {
             if !patterns.is_empty() {
-                // FTS: join patterns with spaces for multi-keyword search
-                let combined_pattern = patterns.join(" ");
-                field_result_sets.push(
-                    query_field_impl(
+                let mut field_union = HashSet::new();
+                for pattern in patterns {
+                    let results = query_field_impl(
                         &lore_table,
                         "symbols".to_string(),
-                        combined_pattern,
+                        pattern.clone(),
                         search_limit,
+                        since_dt,
+                        until_dt,
                     )
-                    .await?,
-                );
+                    .await?;
+                    field_union.extend(results);
+                }
+                field_result_sets.push(field_union);
             }
         }
 
-        // Query recipients field
+        // Query recipients field (OR across patterns)
         if let Some(patterns) = recipients_patterns {
             if !patterns.is_empty() {
-                // FTS: join patterns with spaces for multi-keyword search
-                let combined_pattern = patterns.join(" ");
-                field_result_sets.push(
-                    query_field_impl(
+                let mut field_union = HashSet::new();
+                for pattern in patterns {
+                    let results = query_field_impl(
                         &lore_table,
                         "recipients".to_string(),
-                        combined_pattern,
+                        pattern.clone(),
                         search_limit,
+                        since_dt,
+                        until_dt,
                     )
-                    .await?,
-                );
+                    .await?;
+                    field_union.extend(results);
+                }
+                field_result_sets.push(field_union);
             }
         }
 
@@ -3232,7 +3284,7 @@ impl VectorSearchManager {
         );
 
         // Separate field filters from date filters
-        // Field filters affect which emails to search (FTS/regex), date filters affect final results
+        // Field filters select emails via FTS/regex; date filters narrow candidates during FTS post-filtering
         let has_field_filters = filters.from_patterns.is_some()
             || filters.subject_patterns.is_some()
             || filters.body_patterns.is_some()
@@ -3242,39 +3294,13 @@ impl VectorSearchManager {
         let lore_vectors_table = self.connection.open_table("lore_vectors").execute().await?;
         let lore_table = self.connection.open_table("lore").execute().await?;
 
-        // Build date filter clause if needed
-        let date_filter = match (filters.since_date, filters.until_date) {
-            (Some(since), Some(until)) => {
-                let escaped_since = since.replace("'", "''");
-                let escaped_until = until.replace("'", "''");
-                Some(format!(
-                    "date >= '{}' AND date <= '{}'",
-                    escaped_since, escaped_until
-                ))
-            }
-            (Some(since), None) => {
-                let escaped_since = since.replace("'", "''");
-                Some(format!("date >= '{}'", escaped_since))
-            }
-            (None, Some(until)) => {
-                let escaped_until = until.replace("'", "''");
-                Some(format!("date <= '{}'", escaped_until))
-            }
-            (None, None) => None,
-        };
-
-        if let Some(ref filter) = date_filter {
-            tracing::info!(
-                "vlore: Date filter will be applied during email fetch: {}",
-                filter
-            );
-        }
+        let has_date_filter = filters.since_date.is_some() || filters.until_date.is_some();
 
         // No field filters (but may have date filter): simple vector search
         if !has_field_filters {
             // Note: LanceDB vector search on lore_vectors table doesn't have date column
             // We must fetch more candidates and filter in fetch_emails_by_ids
-            let fetch_multiplier = if date_filter.is_some() {
+            let fetch_multiplier = if has_date_filter {
                 50 // Significantly increase to ensure we get enough results after date filtering
             } else {
                 2
@@ -3352,6 +3378,8 @@ impl VectorSearchManager {
                     filters.symbols_patterns,
                     filters.recipients_patterns,
                     search_limit,
+                    filters.since_date,
+                    filters.until_date,
                 )
                 .await?;
 
@@ -3385,16 +3413,10 @@ impl VectorSearchManager {
             // 3. Build score map from vector results
             let mut score_map = HashMap::new();
             for batch in &vector_results {
-                let message_id_array = batch
-                    .column(0)
-                    .as_any()
-                    .downcast_ref::<arrow::array::StringArray>()
-                    .unwrap();
-                let distance_array = batch
-                    .column(2)
-                    .as_any()
-                    .downcast_ref::<arrow::array::Float32Array>()
-                    .unwrap();
+                let message_id_array: &arrow::array::StringArray =
+                    super::get_column(batch, "message_id")?;
+                let distance_array: &arrow::array::Float32Array =
+                    super::get_column(batch, "_distance")?;
 
                 for i in 0..batch.num_rows() {
                     let message_id = message_id_array.value(i).to_string();
@@ -3431,7 +3453,9 @@ impl VectorSearchManager {
                 continue; // Try larger search if no intersection found
             }
 
-            // 5. Fetch full email data for intersection (only what we need, up to limit)
+            // 5. Fetch full email data for intersection (up to limit).
+            // Date filtering was already applied in the FTS phase,
+            // so no date re-filtering is needed here.
             let ids_to_fetch: Vec<String> = intersection_ids
                 .iter()
                 .take(limit)
@@ -3444,8 +3468,8 @@ impl VectorSearchManager {
                     &score_map,
                     lore_table.clone(),
                     limit,
-                    filters.since_date,
-                    filters.until_date,
+                    None,
+                    None,
                 )
                 .await?;
 
@@ -3527,61 +3551,26 @@ impl VectorSearchManager {
                 .await?;
 
             for batch in &batches {
-                let git_commit_sha_array = batch
-                    .column(0)
-                    .as_any()
-                    .downcast_ref::<arrow::array::StringArray>()
-                    .unwrap();
-                let from_array = batch
-                    .column(1)
-                    .as_any()
-                    .downcast_ref::<arrow::array::StringArray>()
-                    .unwrap();
-                let date_array = batch
-                    .column(2)
-                    .as_any()
-                    .downcast_ref::<arrow::array::StringArray>()
-                    .unwrap();
+                let git_commit_sha_array: &arrow::array::StringArray =
+                    super::get_column(batch, "git_commit_sha")?;
+                let from_array: &arrow::array::StringArray = super::get_column(batch, "from")?;
+                let date_array: &arrow::array::StringArray = super::get_column(batch, "date")?;
                 let date_timestamp_array = batch
-                    .column(3)
-                    .as_any()
-                    .downcast_ref::<arrow::array::Int64Array>()
-                    .unwrap();
-                let message_id_array = batch
-                    .column(4)
-                    .as_any()
-                    .downcast_ref::<arrow::array::StringArray>()
-                    .unwrap();
-                let in_reply_to_array = batch
-                    .column(5)
-                    .as_any()
-                    .downcast_ref::<arrow::array::StringArray>()
-                    .unwrap();
-                let subject_array = batch
-                    .column(6)
-                    .as_any()
-                    .downcast_ref::<arrow::array::StringArray>()
-                    .unwrap();
-                let references_array = batch
-                    .column(7)
-                    .as_any()
-                    .downcast_ref::<arrow::array::StringArray>()
-                    .unwrap();
-                let recipients_array = batch
-                    .column(8)
-                    .as_any()
-                    .downcast_ref::<arrow::array::StringArray>()
-                    .unwrap();
-                let body_array = batch
-                    .column(9)
-                    .as_any()
-                    .downcast_ref::<arrow::array::StringArray>()
-                    .unwrap();
-                let symbols_array = batch
-                    .column(10)
-                    .as_any()
-                    .downcast_ref::<arrow::array::StringArray>()
-                    .unwrap();
+                    .column_by_name("date_timestamp")
+                    .and_then(|c| c.as_any().downcast_ref::<arrow::array::Int64Array>());
+                let message_id_array: &arrow::array::StringArray =
+                    super::get_column(batch, "message_id")?;
+                let in_reply_to_array: &arrow::array::StringArray =
+                    super::get_column(batch, "in_reply_to")?;
+                let subject_array: &arrow::array::StringArray =
+                    super::get_column(batch, "subject")?;
+                let references_array: &arrow::array::StringArray =
+                    super::get_column(batch, "references")?;
+                let recipients_array: &arrow::array::StringArray =
+                    super::get_column(batch, "recipients")?;
+                let body_array: &arrow::array::StringArray = super::get_column(batch, "body")?;
+                let symbols_array: &arrow::array::StringArray =
+                    super::get_column(batch, "symbols")?;
 
                 for i in 0..batch.num_rows() {
                     if results.len() >= limit {
@@ -3642,8 +3631,9 @@ impl VectorSearchManager {
                     let symbols: Vec<String> =
                         serde_json::from_str(symbols_json).unwrap_or_default();
 
-                    // Get date_timestamp from the batch
-                    let date_timestamp = date_timestamp_array.value(i);
+                    let date_timestamp = date_timestamp_array
+                        .map(|arr| arr.value(i))
+                        .unwrap_or_else(|| email_datetime.timestamp());
 
                     results.push((
                         crate::types::LoreEmailInfo {
diff --git a/src/database_utils.rs b/src/database_utils.rs
index 7512c47..c45535d 100644
--- a/src/database_utils.rs
+++ b/src/database_utils.rs
@@ -9,7 +9,9 @@ use std::path::Path;
 /// 1. If `database_arg` is provided:
 ///    - If it's a directory, look for `.semcode.db` within it
 ///    - Otherwise, use the path as-is (direct database path)
-/// 2. If `database_arg` is None:
+/// 2. If `database_arg` is None, check the `SEMCODE_DB` environment variable
+///    (same directory/suffix semantics as the `-d` flag)
+/// 3. If neither is set:
 ///    - For indexing operations: prefer `source_dir/.semcode.db`, fallback to current directory
 ///    - For query operations: use current directory `./.semcode.db`
 ///
@@ -21,23 +23,17 @@ use std::path::Path;
 /// String representation of the database path to use
 pub fn process_database_path(database_arg: Option<&str>, source_dir: Option<&Path>) -> String {
     match database_arg {
-        Some(path) => {
-            let path_obj = Path::new(path);
-
-            // If path already ends with .semcode.db, use it as-is (avoid double appending)
-            if path.ends_with(".semcode.db") {
-                path.to_string()
-            } else if path_obj.is_dir() {
-                // If the path is a directory, look for .semcode.db within it
-                let semcode_db_path = path_obj.join(".semcode.db");
-                semcode_db_path.to_string_lossy().to_string()
-            } else {
-                // If it's a specific file path, use it as-is
-                path.to_string()
-            }
-        }
+        Some(path) => resolve_path(path),
         None => {
-            // No -d flag provided - behavior depends on whether we have a source directory
+            // Check SEMCODE_DB environment variable before falling back to
+            // source-dir or current-dir defaults.
+            if let Ok(env_path) = std::env::var("SEMCODE_DB") {
+                let env_path = env_path.trim();
+                if !env_path.is_empty() {
+                    return resolve_path(env_path);
+                }
+            }
+
             match source_dir {
                 Some(source_path) => {
                     // For indexing operations: prefer source directory unless it's current directory
@@ -58,10 +54,29 @@ pub fn process_database_path(database_arg: Option<&str>, source_dir: Option<&Pat
     }
 }
 
+/// Normalize a database path: append `.semcode.db` to directories, pass
+/// paths that already end with `.semcode.db` through unchanged, and
+/// return anything else as-is.
+fn resolve_path(path: &str) -> String {
+    let path_obj = Path::new(path);
+
+    if path.ends_with(".semcode.db") {
+        path.to_string()
+    } else if path_obj.is_dir() {
+        path_obj.join(".semcode.db").to_string_lossy().to_string()
+    } else {
+        path.to_string()
+    }
+}
+
 #[cfg(test)]
 mod tests {
     use super::*;
     use std::path::Path;
+    use std::sync::Mutex;
+
+    /// Serializes tests that read or write the SEMCODE_DB environment variable.
+    static ENV_LOCK: Mutex<()> = Mutex::new(());
 
     #[test]
     fn test_process_database_path_with_explicit_path() {
@@ -83,24 +98,94 @@ mod tests {
 
     #[test]
     fn test_process_database_path_no_args_no_source() {
-        // Test query mode (no source directory)
+        let _guard = ENV_LOCK.lock().unwrap();
+        let saved = std::env::var("SEMCODE_DB").ok();
+        std::env::remove_var("SEMCODE_DB");
+
         let result = process_database_path(None, None);
         assert_eq!(result, "./.semcode.db");
+
+        if let Some(v) = saved {
+            std::env::set_var("SEMCODE_DB", v);
+        }
     }
 
     #[test]
     fn test_process_database_path_no_args_with_source() {
-        // Test index mode with source directory
+        let _guard = ENV_LOCK.lock().unwrap();
+        let saved = std::env::var("SEMCODE_DB").ok();
+        std::env::remove_var("SEMCODE_DB");
+
         let source_path = Path::new("/source/code");
         let result = process_database_path(None, Some(source_path));
         assert_eq!(result, "/source/code/.semcode.db");
+
+        if let Some(v) = saved {
+            std::env::set_var("SEMCODE_DB", v);
+        }
     }
 
     #[test]
     fn test_process_database_path_current_dir_source() {
-        // Test index mode with current directory as source
+        let _guard = ENV_LOCK.lock().unwrap();
+        let saved = std::env::var("SEMCODE_DB").ok();
+        std::env::remove_var("SEMCODE_DB");
+
         let source_path = Path::new(".");
         let result = process_database_path(None, Some(source_path));
         assert_eq!(result, "./.semcode.db");
+
+        if let Some(v) = saved {
+            std::env::set_var("SEMCODE_DB", v);
+        }
+    }
+
+    #[test]
+    fn test_env_var_used_when_no_flag() {
+        let _guard = ENV_LOCK.lock().unwrap();
+        let saved = std::env::var("SEMCODE_DB").ok();
+        std::env::set_var("SEMCODE_DB", "/data/my-project.semcode.db");
+
+        let result = process_database_path(None, None);
+        assert_eq!(result, "/data/my-project.semcode.db");
+
+        // Also overrides source_dir fallback
+        let result = process_database_path(None, Some(Path::new("/source/code")));
+        assert_eq!(result, "/data/my-project.semcode.db");
+
+        match saved {
+            Some(v) => std::env::set_var("SEMCODE_DB", v),
+            None => std::env::remove_var("SEMCODE_DB"),
+        }
+    }
+
+    #[test]
+    fn test_flag_overrides_env_var() {
+        let _guard = ENV_LOCK.lock().unwrap();
+        let saved = std::env::var("SEMCODE_DB").ok();
+        std::env::set_var("SEMCODE_DB", "/env/path.semcode.db");
+
+        let result = process_database_path(Some("/flag/path.semcode.db"), None);
+        assert_eq!(result, "/flag/path.semcode.db");
+
+        match saved {
+            Some(v) => std::env::set_var("SEMCODE_DB", v),
+            None => std::env::remove_var("SEMCODE_DB"),
+        }
+    }
+
+    #[test]
+    fn test_empty_env_var_ignored() {
+        let _guard = ENV_LOCK.lock().unwrap();
+        let saved = std::env::var("SEMCODE_DB").ok();
+        std::env::set_var("SEMCODE_DB", "");
+
+        let result = process_database_path(None, None);
+        assert_eq!(result, "./.semcode.db");
+
+        match saved {
+            Some(v) => std::env::set_var("SEMCODE_DB", v),
+            None => std::env::remove_var("SEMCODE_DB"),
+        }
     }
 }
diff --git a/src/git.rs b/src/git.rs
index 777bd79..c1b1915 100644
--- a/src/git.rs
+++ b/src/git.rs
@@ -33,6 +33,122 @@ pub fn resolve_to_commit<'a>(repo: &'a gix::Repository, revspec: &str) -> Result
         .try_into_commit()
         .map_err(|_| anyhow::anyhow!("'{}' does not resolve to a commit", revspec))
 }
+
+/// Normalize path string to use forward slashes and trim `$HOME` prefix to `~` if applicable.
+pub fn normalize_repo_path<P: AsRef<Path>>(path: P) -> String {
+    let path = path.as_ref();
+    let canonical = std::fs::canonicalize(path).unwrap_or_else(|_| path.to_path_buf());
+    let home = dirs::home_dir().map(|h| std::fs::canonicalize(&h).unwrap_or(h));
+
+    let path_buf = match home.as_ref().and_then(|h| canonical.strip_prefix(h).ok()) {
+        Some(rel) => PathBuf::from("~").join(rel),
+        None => canonical,
+    };
+
+    path_buf.to_string_lossy().replace('\\', "/")
+}
+
+/// Get canonical repository root identity string (`repo_identity`).
+/// Resolves linked worktrees and bare repos to their shared core `common_dir` path,
+/// with `$HOME` normalized to `~` and backslashes normalized to `/`.
+pub fn get_repo_root<P: AsRef<Path>>(path: P) -> String {
+    let path = path.as_ref();
+    let abs_input = if path.is_absolute() {
+        path.to_path_buf()
+    } else {
+        std::env::current_dir().unwrap_or_default().join(path)
+    };
+
+    let p = gix::discover(&abs_input)
+        .map(|repo| repo.common_dir().to_path_buf())
+        .unwrap_or_else(|_| abs_input.clone());
+
+    let abs_p = if p.is_absolute() {
+        p
+    } else {
+        abs_input.join(p)
+    };
+
+    normalize_repo_path(abs_p)
+}
+
+/// Helper to canonicalize a path or return its original PathBuf if canonicalization fails.
+fn canonicalize_or_original(path: &Path) -> PathBuf {
+    std::fs::canonicalize(path).unwrap_or_else(|_| path.to_path_buf())
+}
+
+/// Canonicalize a path even if the final file or subdirectories do not exist yet on disk.
+fn canonicalize_existing_ancestors(path: &Path) -> PathBuf {
+    if let Ok(canon) = std::fs::canonicalize(path) {
+        return canon;
+    }
+    let mut missing = Vec::new();
+    let mut curr = path;
+    while let Some(parent) = curr.parent() {
+        if let Some(name) = curr.file_name() {
+            missing.push(name);
+        }
+        if let Ok(canon_parent) = std::fs::canonicalize(parent) {
+            let mut result = canon_parent;
+            for component in missing.into_iter().rev() {
+                result.push(component);
+            }
+            return result;
+        }
+        curr = parent;
+    }
+    path.to_path_buf()
+}
+
+/// Determine if a database path (`db_path`) is local to a repository (`repo_path`).
+/// Returns `(repo_identity, is_local_db)`.
+pub fn determine_db_locality<P: AsRef<Path>>(repo_path: P, db_path: &str) -> (String, bool) {
+    let repo_path = repo_path.as_ref();
+    let git_repo_identity = get_repo_root(repo_path);
+    let mut local_roots = Vec::new();
+
+    if let Ok(repo) = gix::discover(repo_path) {
+        let common = canonicalize_or_original(repo.common_dir());
+        if !repo.is_bare() && common.ends_with(".git") {
+            if let Some(main_root) = common.parent() {
+                local_roots.push(main_root.to_path_buf());
+            }
+        }
+        local_roots.push(common);
+
+        if let Some(workdir) = repo.workdir() {
+            local_roots.push(canonicalize_or_original(workdir));
+        }
+    } else {
+        local_roots.push(canonicalize_or_original(repo_path));
+    }
+
+    let db_path_buf = PathBuf::from(db_path);
+    let preferred_root = local_roots.iter().rev().find(|p| p.is_dir());
+
+    let db_abs = if db_path_buf.is_absolute() {
+        db_path_buf
+    } else if let Some(target_root) = preferred_root {
+        target_root.join(db_path_buf)
+    } else {
+        std::env::current_dir()
+            .unwrap_or_default()
+            .join(db_path_buf)
+    };
+
+    let real_target = std::fs::read_link(&db_abs).unwrap_or_else(|_| db_abs.clone());
+    let real_target_abs = if real_target.is_absolute() {
+        real_target
+    } else {
+        db_abs.parent().unwrap_or(Path::new(".")).join(real_target)
+    };
+
+    let db_canonical = canonicalize_existing_ancestors(&real_target_abs);
+    let is_local_db = local_roots.iter().any(|r| db_canonical.starts_with(r));
+
+    (git_repo_identity, is_local_db)
+}
+
 use once_cell::sync::Lazy;
 use std::collections::HashSet;
 use std::path::{Path, PathBuf};
@@ -223,6 +339,146 @@ pub fn get_git_file_hash_with_fallback<P: AsRef<Path>>(file_path: P) -> Result<S
 mod tests {
     use super::*;
 
+    #[test]
+    fn test_normalize_repo_path() {
+        let current_dir = std::env::current_dir().unwrap();
+        let normalized = normalize_repo_path(&current_dir);
+        assert!(!normalized.contains('\\'));
+        if let Some(home) = dirs::home_dir() {
+            if current_dir.starts_with(&home) {
+                assert!(normalized.starts_with("~"));
+            }
+        }
+    }
+
+    #[test]
+    fn test_get_repo_root() {
+        let root = get_repo_root(".");
+        assert!(!root.is_empty());
+        assert!(!root.contains('\\'));
+        if let Some(home) = dirs::home_dir() {
+            if std::env::current_dir().unwrap().starts_with(&home) {
+                assert!(root.starts_with("~"));
+            }
+        }
+    }
+
+    #[test]
+    fn test_determine_db_locality_local() {
+        let (identity, is_local) = determine_db_locality(".", ".semcode.db");
+        assert!(!identity.is_empty());
+        assert!(is_local, "Database inside repository should be local");
+    }
+
+    #[test]
+    fn test_determine_db_locality_external() {
+        let (identity, is_local) = determine_db_locality(".", "/tmp/semcode_external_test.db");
+        assert!(!identity.is_empty());
+        assert!(
+            !is_local,
+            "Database outside repository should be shared/external"
+        );
+    }
+
+    #[test]
+    fn test_determine_db_locality_non_existent_nested_path() {
+        let (identity, is_local) =
+            determine_db_locality(".", "target/non_existent_subdir/deep_nested/db.semcode");
+        assert!(!identity.is_empty());
+        assert!(
+            is_local,
+            "Relative DB path inside non-existent subdirectory of repo should be local"
+        );
+    }
+
+    #[test]
+    fn test_determine_db_locality_bare_repo() {
+        let temp_dir = tempfile::tempdir().unwrap();
+        let bare_repo_dir = temp_dir.path().join("test_bare.git");
+        let status = std::process::Command::new("git")
+            .args(["init", "--bare", bare_repo_dir.to_str().unwrap()])
+            .env("GIT_CONFIG_GLOBAL", "/dev/null")
+            .env("GIT_CONFIG_SYSTEM", "/dev/null")
+            .status()
+            .unwrap();
+        assert!(status.success());
+
+        // Local DB inside bare repo directory
+        let local_db = bare_repo_dir.join("semcode.db");
+        let (identity, is_local) =
+            determine_db_locality(&bare_repo_dir, local_db.to_str().unwrap());
+        assert!(!identity.is_empty());
+        assert!(is_local, "DB inside bare repo directory should be local");
+
+        // External DB outside bare repo
+        let external_db = temp_dir.path().join("external.db");
+        let (_, is_local_ext) =
+            determine_db_locality(&bare_repo_dir, external_db.to_str().unwrap());
+        assert!(
+            !is_local_ext,
+            "DB outside bare repo directory should be external"
+        );
+    }
+
+    #[test]
+    fn test_determine_db_locality_worktree() {
+        let temp_dir = tempfile::tempdir().unwrap();
+        let main_repo = temp_dir.path().join("main_repo");
+        std::fs::create_dir_all(&main_repo).unwrap();
+
+        // Init main repo and create initial commit
+        let run_git = |dir: &Path, args: &[&str]| {
+            let res = std::process::Command::new("git")
+                .args(args)
+                .current_dir(dir)
+                .env("GIT_AUTHOR_NAME", "test")
+                .env("GIT_AUTHOR_EMAIL", "test@test.com")
+                .env("GIT_COMMITTER_NAME", "test")
+                .env("GIT_COMMITTER_EMAIL", "test@test.com")
+                .env("GIT_CONFIG_GLOBAL", "/dev/null")
+                .env("GIT_CONFIG_SYSTEM", "/dev/null")
+                .output()
+                .unwrap();
+            assert!(res.status.success(), "Git command failed: {:?}", res);
+        };
+
+        run_git(&main_repo, &["init", "-b", "main"]);
+        std::fs::write(main_repo.join("README.md"), "hello").unwrap();
+        run_git(&main_repo, &["add", "."]);
+        run_git(&main_repo, &["commit", "-m", "initial commit"]);
+
+        // Create linked worktree
+        let worktree_dir = temp_dir.path().join("linked_worktree");
+        run_git(
+            &main_repo,
+            &[
+                "worktree",
+                "add",
+                worktree_dir.to_str().unwrap(),
+                "-b",
+                "wt-branch",
+            ],
+        );
+
+        // DB inside linked worktree
+        let wt_db = worktree_dir.join(".semcode.db");
+        let (wt_identity, is_wt_local) =
+            determine_db_locality(&worktree_dir, wt_db.to_str().unwrap());
+        assert!(is_wt_local, "DB inside linked worktree should be local");
+
+        // DB inside main repo queried from linked worktree
+        let main_db = main_repo.join(".semcode.db");
+        let (main_identity, is_main_local) =
+            determine_db_locality(&worktree_dir, main_db.to_str().unwrap());
+        assert!(
+            is_main_local,
+            "DB inside main repo queried from worktree should be local"
+        );
+
+        // Identity of worktree and main repo should match (canonical common_dir)
+        assert_eq!(wt_identity, main_identity);
+    }
+
     #[test]
     fn test_get_git_sha_current_dir() {
         // This should work if running in a git repository
diff --git a/src/indexer.rs b/src/indexer.rs
index c32f820..a9fe275 100644
--- a/src/indexer.rs
+++ b/src/indexer.rs
@@ -5,6 +5,7 @@
 //! that can be called from different binaries (semcode-index, semcode, semcode-mcp, etc.)
 
 use anyhow::Result;
+use gix::revision::walk::Sorting;
 use std::collections::HashSet;
 use std::path::PathBuf;
 use std::sync::atomic::{AtomicUsize, Ordering};
@@ -87,12 +88,11 @@ pub async fn check_and_optimize_if_needed(
 
 /// Parse git range and get all commit SHAs in the range.
 ///
-/// Walks first-parent chain from B back to A, avoiding gix's rev_walk + with_hidden
-/// which does an O(all commits) graph painting that is catastrophically slow on large repos.
+/// List commit SHAs in a range using gitoxide's rev_walk.
 ///
 /// Supports two formats:
-/// 1. "A..B" - commits reachable from B but not from A (first-parent walk)
-/// 2. "REF" (no ..) - all commits reachable from REF via first-parent (for initial indexing)
+/// 1. "A..B" - commits reachable from B but not from A
+/// 2. "REF" (no ..) - all commits reachable from REF (for initial indexing)
 pub fn list_shas_in_range(repo: &gix::Repository, range: &str) -> Result<Vec<String>> {
     let (from_spec, to_spec) = if range.contains("..") {
         let parts: Vec<&str> = range.split("..").collect();
@@ -107,14 +107,20 @@ pub fn list_shas_in_range(repo: &gix::Repository, range: &str) -> Result<Vec<Str
     let to_commit = resolve_to_commit(repo, to_spec)?;
     let to_id = to_commit.id().detach();
 
-    let stop_at = if from_spec.is_empty() {
-        None
+    let walk = if from_spec.is_empty() {
+        repo.rev_walk([to_id])
+            .sorting(Sorting::ByCommitTime(Default::default()))
+            .all()?
     } else {
         match resolve_to_commit(repo, from_spec) {
-            Ok(from_commit) => Some(from_commit.id().detach()),
+            Ok(from_commit) => {
+                let from_id = from_commit.id().detach();
+                repo.rev_walk([to_id])
+                    .with_hidden([from_id])
+                    .sorting(Sorting::ByCommitTime(Default::default()))
+                    .all()?
+            }
             Err(e) => {
-                // Parent commit not found (shallow clone or root commit).
-                // Fall back to returning just the target commit.
                 info!(
                     "Could not resolve '{}' ({}); falling back to single-commit indexing",
                     from_spec, e
@@ -125,18 +131,10 @@ pub fn list_shas_in_range(repo: &gix::Repository, range: &str) -> Result<Vec<Str
     };
 
     let mut shas = Vec::new();
-    let mut current_id = to_id;
     const MAX_COMMITS: usize = 10000000;
 
-    loop {
-        // Stop if we've reached the exclusion boundary
-        if let Some(ref stop) = stop_at {
-            if current_id == *stop {
-                break;
-            }
-        }
-
-        shas.push(current_id.to_string());
+    for info in walk {
+        let info = info?;
 
         if shas.len() > MAX_COMMITS {
             return Err(anyhow::anyhow!(
@@ -146,20 +144,7 @@ pub fn list_shas_in_range(repo: &gix::Repository, range: &str) -> Result<Vec<Str
             ));
         }
 
-        // Get the first parent to continue walking.
-        // In shallow clones, parent objects may be missing — treat as end of history.
-        let commit = match repo.find_commit(current_id) {
-            Ok(c) => c,
-            Err(_) => break,
-        };
-        let parent_ids: Vec<_> = commit.parent_ids().collect();
-
-        if parent_ids.is_empty() {
-            // Root commit, no more parents
-            break;
-        }
-
-        current_id = parent_ids[0].detach();
+        shas.push(info.id().to_string());
     }
 
     // Reverse to get chronological order (oldest first)
@@ -985,42 +970,57 @@ pub async fn process_lore_commits_pipeline(
 
                 match batch {
                     Ok(emails) => {
-                        let batch_len = emails.len();
-
                         // Insert batch into database
-                        if let Err(e) = db_manager_clone.insert_lore_emails(&emails).await {
-                            error!("Inserter {} failed to insert batch: {}", inserter_id, e);
-                        } else {
-                            // Record processed commit SHAs so they are
-                            // not re-examined on subsequent runs.
-                            let shas: Vec<String> = emails
-                                .iter()
-                                .map(|e| e.git_commit_sha.as_str())
-                                .collect::<std::collections::HashSet<_>>()
-                                .into_iter()
-                                .map(String::from)
-                                .collect();
-                            if let Err(e) =
-                                db_manager_clone.insert_lore_indexed_commits(&shas).await
-                            {
-                                error!(
-                                    "Inserter {} failed to record indexed commits: {}",
-                                    inserter_id, e
-                                );
+                        match db_manager_clone.insert_lore_emails(&emails).await {
+                            Err(e) => {
+                                error!("Inserter {} failed to insert batch: {}", inserter_id, e);
                             }
+                            Ok(failed_indices) => {
+                                let failed_set: std::collections::HashSet<usize> =
+                                    failed_indices.into_iter().collect();
+                                let success_count = emails.len() - failed_set.len();
+
+                                // Record commit SHAs only for emails that
+                                // were actually stored, so that failed
+                                // emails are retried on the next run.
+                                let shas: Vec<String> = emails
+                                    .iter()
+                                    .enumerate()
+                                    .filter(|(i, _)| !failed_set.contains(i))
+                                    .map(|(_, e)| e.git_commit_sha.as_str())
+                                    .collect::<std::collections::HashSet<_>>()
+                                    .into_iter()
+                                    .map(String::from)
+                                    .collect();
+                                if !shas.is_empty() {
+                                    if let Err(e) =
+                                        db_manager_clone.insert_lore_indexed_commits(&shas).await
+                                    {
+                                        error!(
+                                            "Inserter {} failed to record indexed commits: {}",
+                                            inserter_id, e
+                                        );
+                                    }
+                                }
 
-                            let count = inserted_clone.fetch_add(batch_len, Ordering::Relaxed);
-                            pb_clone.set_message(format!("Inserted {} emails", count + batch_len));
-
-                            // Track successful batch insertions and check for periodic optimization
-                            let total_batches = batches_counter.fetch_add(1, Ordering::Relaxed) + 1;
-                            check_and_optimize_if_needed(
-                                &db_manager_clone,
-                                inserter_id,
-                                total_batches,
-                                &optimization_check_timer,
-                            )
-                            .await;
+                                let count =
+                                    inserted_clone.fetch_add(success_count, Ordering::Relaxed);
+                                pb_clone.set_message(format!(
+                                    "Inserted {} emails",
+                                    count + success_count
+                                ));
+
+                                // Track successful batch insertions and check for periodic optimization
+                                let total_batches =
+                                    batches_counter.fetch_add(1, Ordering::Relaxed) + 1;
+                                check_and_optimize_if_needed(
+                                    &db_manager_clone,
+                                    inserter_id,
+                                    total_batches,
+                                    &optimization_check_timer,
+                                )
+                                .await;
+                            }
                         }
                     }
                     Err(_) => {
@@ -1136,3 +1136,153 @@ pub async fn index_git_commits(
 
     Ok(commit_count)
 }
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    /// Helper: run a git command in a directory, panic on failure.
+    fn git(repo: &std::path::Path, args: &[&str]) -> String {
+        let out = std::process::Command::new("git")
+            .args(args)
+            .current_dir(repo)
+            .env("GIT_AUTHOR_NAME", "test")
+            .env("GIT_AUTHOR_EMAIL", "test@test.com")
+            .env("GIT_COMMITTER_NAME", "test")
+            .env("GIT_COMMITTER_EMAIL", "test@test.com")
+            .env("GIT_CONFIG_GLOBAL", "/dev/null")
+            .env("GIT_CONFIG_SYSTEM", "/dev/null")
+            .output()
+            .expect("git command failed to execute");
+        assert!(
+            out.status.success(),
+            "git {:?} failed: {}",
+            args,
+            String::from_utf8_lossy(&out.stderr)
+        );
+        String::from_utf8_lossy(&out.stdout).trim().to_string()
+    }
+
+    /// Build a repo with merge topology and a tag on the side branch:
+    ///
+    ///   A -- B -- M -- D   (main)
+    ///         \      /
+    ///          C ---       (tagged "v1")
+    ///
+    /// Returns (tempdir, repo, sha_a, sha_b, sha_c, sha_m, sha_d).
+    fn create_merge_repo() -> (
+        tempfile::TempDir,
+        gix::Repository,
+        String,
+        String,
+        String,
+        String,
+        String,
+    ) {
+        let tmpdir = tempfile::tempdir().unwrap();
+        let p = tmpdir.path();
+
+        git(p, &["init", "-b", "main"]);
+
+        // Commit A
+        std::fs::write(p.join("a.txt"), "a").unwrap();
+        git(p, &["add", "a.txt"]);
+        git(p, &["commit", "-m", "A"]);
+        let sha_a = git(p, &["rev-parse", "HEAD"]);
+
+        // Commit B
+        std::fs::write(p.join("b.txt"), "b").unwrap();
+        git(p, &["add", "b.txt"]);
+        git(p, &["commit", "-m", "B"]);
+        let sha_b = git(p, &["rev-parse", "HEAD"]);
+
+        // Side branch with commit C, tagged "v1"
+        git(p, &["checkout", "-b", "side"]);
+        std::fs::write(p.join("c.txt"), "c").unwrap();
+        git(p, &["add", "c.txt"]);
+        git(p, &["commit", "-m", "C"]);
+        let sha_c = git(p, &["rev-parse", "HEAD"]);
+        git(p, &["tag", "v1"]);
+
+        // Back to main, merge side → creates merge commit M
+        git(p, &["checkout", "main"]);
+        git(p, &["merge", "side", "--no-ff", "-m", "M"]);
+        let sha_m = git(p, &["rev-parse", "HEAD"]);
+
+        // Commit D on top
+        std::fs::write(p.join("d.txt"), "d").unwrap();
+        git(p, &["add", "d.txt"]);
+        git(p, &["commit", "-m", "D"]);
+        let sha_d = git(p, &["rev-parse", "HEAD"]);
+
+        let repo = gix::open(p).unwrap();
+        (tmpdir, repo, sha_a, sha_b, sha_c, sha_m, sha_d)
+    }
+
+    /// Regression test: a tag on a side branch (not on the first-parent
+    /// chain) must still correctly bound the range.  A first-parent-only
+    /// walk would miss it and traverse the entire history.
+    #[test]
+    fn test_list_shas_tag_on_side_branch() {
+        let (_tmp, repo, _sha_a, _sha_b, _sha_c, sha_m, sha_d) = create_merge_repo();
+
+        let shas = list_shas_in_range(&repo, "v1..HEAD").unwrap();
+
+        // v1 points at C (on the side branch).  The range v1..HEAD
+        // should include M and D only — NOT A, B, or C.
+        assert!(
+            shas.contains(&sha_m),
+            "merge commit M should be in v1..HEAD"
+        );
+        assert!(shas.contains(&sha_d), "commit D should be in v1..HEAD");
+        assert_eq!(
+            shas.len(),
+            2,
+            "v1..HEAD should contain exactly 2 commits (M, D), got {}: {:?}",
+            shas.len(),
+            shas
+        );
+    }
+
+    /// Range where the start IS on the first-parent chain works too.
+    #[test]
+    fn test_list_shas_first_parent_boundary() {
+        let (_tmp, repo, sha_a, sha_b, sha_c, sha_m, sha_d) = create_merge_repo();
+
+        let shas = list_shas_in_range(&repo, &format!("{sha_a}..HEAD")).unwrap();
+
+        // A..HEAD should include B, C, M, D (everything after A).
+        let expected: std::collections::HashSet<&str> = [
+            sha_b.as_str(),
+            sha_c.as_str(),
+            sha_m.as_str(),
+            sha_d.as_str(),
+        ]
+        .into_iter()
+        .collect();
+        let got: std::collections::HashSet<&str> = shas.iter().map(|s| s.as_str()).collect();
+
+        assert_eq!(expected, got, "A..HEAD should be {{B, C, M, D}}");
+    }
+
+    /// No range separator — returns all reachable commits.
+    #[test]
+    fn test_list_shas_no_range() {
+        let (_tmp, repo, sha_a, sha_b, sha_c, sha_m, sha_d) = create_merge_repo();
+
+        let shas = list_shas_in_range(&repo, "HEAD").unwrap();
+
+        let expected: std::collections::HashSet<&str> = [
+            sha_a.as_str(),
+            sha_b.as_str(),
+            sha_c.as_str(),
+            sha_m.as_str(),
+            sha_d.as_str(),
+        ]
+        .into_iter()
+        .collect();
+        let got: std::collections::HashSet<&str> = shas.iter().map(|s| s.as_str()).collect();
+
+        assert_eq!(expected, got, "HEAD (no range) should return all 5 commits");
+    }
+}