diff --git a/CLAUDE.md b/CLAUDE.md
index 9ccde1f..bc2c944 100644
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -53,7 +53,7 @@ Two main interfaces built on NUI.nvim:
 - **Type System**: Extensive Lua type annotations in `types.lua` for IDE support
 - **Zero Config**: Works out of the box, optional setup in user config
 - **Dependencies**: Requires `nui.nvim`, `telescope.nvim`, and `curl`
-- **Environment**: Set `OPENAI_API_KEY` for OpenAI, requires local Ollama server for Ollama provider
+- **Environment**: Set `OPENAI_API_KEY` for OpenAI, requires Ollama server for Ollama provider (configurable via `vim.g.gptmodels_ollama_host`)
 - **Visual Selection**: Commands automatically detect and handle visual mode selections
 - **Background Processing**: Windows can close while maintaining streaming responses
 
diff --git a/README.md b/README.md
index b5ed3be..3eae678 100644
--- a/README.md
+++ b/README.md
@@ -91,6 +91,7 @@ them here. But here they are anyways:
 
 This plugin requires `curl` be installed for requests.
 For Ollama requests, have [Ollama](https://ollama.com/) running locally.
+To use a remote Ollama server, set `vim.g.gptmodels_ollama_host` to the server's URL (e.g., `http://remote-host:11434`).
 For OpenAI requests, have the `OPENAI_API_KEY` environment variable set.
 
 Now, in your favorite package manager:
diff --git a/TODO.md b/TODO.md
index c618092..4ab66e4 100644
--- a/TODO.md
+++ b/TODO.md
@@ -3,7 +3,6 @@
 * Make sure o1 models are available
 * Let some ops, like attach files, work in insert mode as well
 * Let <CR> work from all windows
-* Offer some config for ollama location
 * chat window's prompt needs to be adjusted so the user request counts as much as the included files,
   currently it's dominated by the files and many llms end up just explaining what's in the files
 * Get live reloading of C-p telescope model picker when openai results come in