davsclaus opened a new pull request, #26144: URL: https://github.com/apache/camel/pull/26144
## Summary - **`camel doctor`**: detects Ollama at `localhost:11434`, lists pulled models. Shows a friendly hint when not running (optional, for local AI). - **TUI DoctorPopup**: new Ollama row with model count and names; warns when all pulled models are <14B parameters (too small for reliable tool calling in F8); no AI-provider warning when Ollama is running. - **`LlmClient`**: reads `LLM_BASE_URL` (or `OPENAI_BASE_URL`) env var so `LLM_API_KEY` can point at any OpenAI-compatible local server (LM Studio, vLLM, llama.cpp, GPT4All). - **AiPanel F8 setup guide**: shown when no LLM provider is reachable — includes Ollama install steps (native only, not Docker), ≥14B model table with RAM requirements, tool-calling warning, cloud provider table, and OpenAI-compatible server section. - **Docs** (`camel-jbang-ai.adoc`, `camel-jbang-tui.adoc`): Ollama native install, ≥14B model requirements, tool-calling warning, OpenAI-compatible server section. Small models (<14B) removed from all guides. ## Motivation The F8 TUI AI panel and `camel ask` use tool calling to inspect live Camel processes. Models smaller than ~14B do not reliably invoke tools and answer from training knowledge instead. This PR adds first-class Ollama support with clear guidance on model selection, and makes `LLM_API_KEY` useful with any OpenAI-compatible server via `LLM_BASE_URL`. Jira: https://issues.apache.org/jira/browse/CAMEL-24631 ## Test plan - [ ] `camel doctor` with Ollama running — shows model list - [ ] `camel doctor` with Ollama not running — shows optional hint - [ ] TUI Doctor popup (F8 → Doctor) with Ollama running — shows model row - [ ] TUI Doctor popup with only small models (e.g. `:8b`) — shows WARN + upgrade hint - [ ] TUI F8 with no provider — setup guide renders with install instructions - [ ] `LLM_API_KEY=x LLM_BASE_URL=http://localhost:1234 camel ask` — routes to custom server 🤖 Generated with [Claude Code](https://claude.ai/claude-code) _Claude Code on behalf of davsclaus_ -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
