Concepts

Embedded intent inference

JoyMux ships a self-contained local understanding layer. A normal install must not require Ollama, llama.cpp, ONNX Runtime, Python, Node.js, model servers, or manual model downloads before basic natural-language understanding works.

Guarantee

JoyMux production startup must not depend on Ollama.

Ollama and OpenAI-compatible local endpoints may exist as development-only backends. They are never the production default and must not appear in normal onboarding.

Boundary

The embedded model is not a coding harness. It does not install software, execute shell commands, select unsafe commands, grant permissions, choose a coding harness, or own JoyMesh routing.

JoyMux never selects an AI reasoning harness. JoyMesh / external clients attach. The understanding layer produces a typed intent; policy, approval, and execution stay outside that layer. Model confidence must never skip approval.

What you run

joymux ask "Install OpenClaw for me."
joymux model status|verify|repair|info
joymux doctor

Normal users are not required to run model or doctor commands during successful onboarding. The UI/CLI may open immediately; obvious requests are handled while the local model becomes ready.

If the local model is missing, corrupted, or unavailable on this platform, joymux doctor and joymux model status report that. Offline use keeps the last-known-good path.