Feature cleanup #12
Owner
- Backend — LAN counts as local. is_local_endpoint now accepts loopback and private LAN — 192.168.x, 10.x, 172.16–31.x, 169.254.x, *.local, plus IPv6 ULA/link-local. So a self-hosted Ollama on another box on your network is treated like localhost — no "data leaves your network" warning, no allowlist flag. Public/internet hosts still are. (Unit-tested: 192.168.0.42 ✅ local, 8.8.8.8 ✅ not.)
- API fix — setLlmProvider was sending { config }; the command expects { args }. Fixed (it was never actually wired).
- New "AI" settings tab — pick provider (Off / Ollama / Custom OpenAI-compatible), set the endpoint (http://192.168.0.42:11434) and model, and Test / refresh to see reachability + the model list. A live banner warns only if you point it at something outside your LAN.
1. Backend — LAN counts as local. is_local_endpoint now accepts loopback and private LAN — 192.168.x, 10.x, 172.16–31.x, 169.254.x, *.local, plus IPv6 ULA/link-local. So a self-hosted Ollama on another box on your network is treated like localhost — no "data leaves your network" warning, no allowlist flag. Public/internet hosts still are. (Unit-tested: 192.168.0.42 ✅ local, 8.8.8.8 ✅ not.)
2. API fix — setLlmProvider was sending { config }; the command expects { args }. Fixed (it was never actually wired).
3. New "AI" settings tab — pick provider (Off / Ollama / Custom OpenAI-compatible), set the endpoint (http://192.168.0.42:11434) and model, and Test / refresh to see reachability + the model list. A live banner warns only if you point it at something outside your LAN.
iamdoubz
added 14 commits 2026-07-03 11:42:29 -05:00
Reference in New Issue
Block a user