Local models via localhost:11434
Ollama-Pippa runs entirely on the local machine. No API calls leave the machine. Useful for scheduled background tasks where Claude's degraded service window would otherwise drop the job, and for offline situations.
Native /api/chat
Streams via Ollama's native /api/chat endpoint. Multi-turn tool loop with {type, function: {name, ...}} schema wrapping (Ollama's quirk — flat OpenAI-style schema gets rejected) and tool call ID propagation through the loop.
Sticky model selection
Active model is persisted to POST /api/ollama/model. GET /api/ollama/models lists installed local models. Switching is one click in the brain switcher.
Tip: Ollama is the safety net for the heartbeat (Heart track, lesson 5). When Claude is degraded and Codex/Gemini are also unavailable, scheduled tasks still run — slower, smaller, but they run. That matters for daily-7am-weather kinds of jobs.