ifURI App

AI Cost Tracking

• 🤖 LLM usage: $0.8012 (5 commits)
• 👤 Human dev: ~$361 (3.6h @ $100/h, 30min dedup)

Generated on 2026-06-17 using openrouter/qwen/qwen3-coder-next

---

Desktop and browser client for IFURI — voice commands, multi-endpoint chat, URI flow execution, LAN pairing and optional urirun registry execution.

Published at github.com/if-uri/app · ifuri.com

What it does

ifURI (ifuri-app) lets users:

• chat with each urisys-node :8790, MCP, A2A, LLM, ifURI peer and WebRTC peer in /voice,
• keep URL state (lang, theme, view, channel, prompt) in shareable links,
• store chat history on urisys-node (/app/chat/*) or locally as fallback,
• run URI flows from urisys-examples,
• call local or service-backed urirun registries through ifuri-app urirun-call and /api/urirun/call,
• use the Tkinter desktop app (flows + LAN + czaty).

Autonomy (koru + twin-human)

Aplikacja integruje się z koru autonomous loop + urirun-twin-human:

• Zadania planfile z labelami kvm, lenovo, signal-gui (np. IFURI-226: wysyłka Signal na desktopie lenovo) są automatycznie delegowane do twin-human.
• Realne komendy URI (kvm://laptop/...) są wykonywane i logowane.
• Uruchomienie: urirun start lub make koru-cycle.
• Logi widoczne w dashboardzie "Na żywo — koru (realne komendy URI)".
• Wsparcie dla Digital Twin (osoby + grants + unblock_ledger).

Makefile (recommended)

cd ~/github/if-uri/app
make help              # lista komend
make install-dev       # editable + pytest
make test              # pytest (unit + API)
make run-voice         # http://127.0.0.1:8766/voice
make run-gui           # desktop Tkinter
make api-smoke         # curl health + chat endpoints
make chat-status URISYS=http://192.168.188.201:8790
make chat-migrate-dry URISYS=http://192.168.188.201:8790   # po upgrade node
make webrtc-install-pack URISYS=http://192.168.188.201:8790
make webrtc-smoke URISYS=http://192.168.188.201:8790
make run ARGS="urirun-info"

Koru autonomy + twin-human (kvm / lenovo desktop)

make koru-cycle          # uruchom cykl z apply (używa twin-human dla ticketów kvm/lenovo)
make koru-plan           # dry-run plan pętli
make koru-execute-twin   # bezpośrednie wywołanie twin-human dla IFURI-226 itp.
make koru-logs           # tail .planfile/.koru/queue.log (to co widać w panelu "Na żywo")
make koru-status         # stan koru + otwarte tickety

Zmienne: PORT=8766, URISYS=http://192.168.188.201:8790, PYTHON=python3

Tło:

make run-voice-bg URISYS=http://192.168.188.201:8790
make health PORT=8766
make stop

Install

Pakiet PyPI: ifuri (aktualnie 0.2.10).

Developer install. make install-dev użyje uv, jeśli jest dostępny; w trybie bez uv instaluje aplikację editable i opcjonalnie lokalne checkouty uri2flow / uricore, jeśli istnieją jako sąsiednie repozytoria developerskie.

make install-dev
# bez uv, minimalnie:
python -m pip install -e ".[flows,dev,packs]"

cd ~/github/if-uri/app
make install-dev
# lub: python -m pip install -e ".[flows]"
ifuri-app --version

Optional urirun runtime:

python -m pip install -e ".[urirun]"
ifuri-app urirun-info
ifuri-app urirun-call tool://local/report/render \
  --registry generated/registry.json \
  --payload '{"format":"html"}'

Quick start

# Na maszynie z urisys-node (np. lenovo)
urisys-node serve --host 0.0.0.0 --port 8790

# Operator — browser
export URISYS_NODE_ENDPOINT=http://192.168.188.201:8790
make run-voice URISYS=http://192.168.188.201:8790
# → http://127.0.0.1:8766/voice?lang=pl&theme=dark

# Operator — native shell (Rust + Tauri, dev)
make run-tauri-dev URISYS=http://192.168.188.201:8790

CLI

ifuri-app app                         # Tkinter desktop
ifuri-app init --scan-lan             # create workspace, discover urisys-node on /24
ifuri-app serve                       # start local runtime HTTP API (no voice UI)
ifuri-app voice --prompt "health"     # voice UI + URL z prompt=
ifuri-app node-health --endpoint http://192.168.188.201:8790
ifuri-app node-call uri://... --endpoint http://192.168.188.201:8790
ifuri-app node-control-test --endpoint http://192.168.188.201:8790
ifuri-app node-screen --endpoint http://192.168.188.201:8790 --out screen.png
ifuri-app flow-run 08-kvm-linkedin.uri.flow.yaml --endpoint http://192.168.188.201:8790
ifuri-app chat-channels
ifuri-app chat-send "status" --endpoint http://192.168.188.201:8790
ifuri-app chat-status --endpoint http://192.168.188.201:8790
ifuri-app chat-migrate --endpoint http://192.168.188.201:8790
ifuri-app packs
ifuri-app run tool://local/report/render --payload '{"format":"html"}' --dry-run
ifuri-app run tool://local/report/render --payload '{"format":"html"}' --execute
ifuri-app expand lenovo-remote/01-health-probe.uri.flow.yaml
ifuri-app urirun-info
ifuri-app urirun-call tool://local/report/render --registry generated/registry.json --payload '{"format":"html"}'
ifuri-app urirun-scan . --out generated/bindings.json
ifuri-app urirun-serve --registry generated/registry.json   # HTTP /health /routes POST /run
ifuri-app urirun-mcp tools --registry generated/registry.json   # tools|card|serve (MCP / A2A)
ifuri-app flow-validate lenovo-remote/01-health-probe.uri.flow.yaml
ifuri-app voice-plan "sprawdź health"
ifuri-app voice-catalog
ifuri-app voice-run "sprawdź health" --dry-run
ifuri-app voice-capabilities --endpoint http://192.168.188.201:8790
ifuri-app voice-install-packs --endpoint http://192.168.188.201:8790
ifuri-app webrtc-capabilities --endpoint http://192.168.188.201:8790
ifuri-app webrtc-install-pack --endpoint http://192.168.188.201:8790
ifuri-app webrtc-smoke --endpoint http://192.168.188.201:8790
ifuri-app discover

Autonomy & Koru (new)

urirun start                 # start koru autonomous loop (twin-human for kvm/lenovo)
urirun start --apply

• Zadania z labels: kvm,lenovo,signal-gui (np. wysyłka Signal na desktopie Lenovo) są obsługiwane przez urirun-twin-human.
• Rzeczywiste komendy URI (kvm://laptop/...) są wykonywane i logowane.
• Panel "Na żywo — koru (realne komendy URI)" pokazuje co naprawdę leci na węzeł.
• Make targets: koru-cycle, koru-plan, koru-logs, koru-status.

Runtime API

Pełna dokumentacja: docs/API.md · diagram: docs/ARCHITECTURE.md · WebRTC: docs/WEBRTC.md

Method	Path	Purpose
GET	/voice	Browser chat UI
GET	/api/health	ifURI + urisys health
GET	/api/chat/channels	LAN endpoints as chats
GET	/api/chat/history?channel_id=	Thread history
POST	/api/chat/send	Send message (text or prompt)
POST	/api/voice/run	Voice pipeline
GET	/api/webrtc/signal	WebRTC signaling inbox (poll)
POST	/api/webrtc/signal	WebRTC SDP/ICE relay
GET	/api/urirun	Optional urirun runtime status
POST	/api/urirun/call	Call URI through urirun registry
POST	/api/urisys/call	Proxy to node

Data

~/.ifuri/workspace.json      # flows, services, urisys endpoint
~/.ifuri/app-chat.jsonl      # chat fallback (gdy node bez /app/chat)

Override: IFURI_HOME, IFURI_CHAT_STORE

Backlog

TODO.md · CHANGELOG.md

Related ifURI projects

• Runtime: if-uri/urirun (w tym urirun start, twin-human, host work/koru)
• Koru autonomy + loop: urirun-connector-loop + urirun-connector-work
• Twin for desktop/KVM: urirun-twin-human
• Public docs: if-uri/docs
• Examples and Docker/noVNC flows: if-uri/examples
• Connector hub: connect.ifuri.com
• Installer: get.ifuri.com
• Current cross-repository summary: work-summary-2026-06-20

License

Licensed under Apache-2.0.