API Reference

Note: This is a quick reference of common endpoints. For the complete API reference, see docs/API.md in the repository.

OpenAI Compatible

Method	Endpoint	Description
GET	/v1/models	List available models
POST	/v1/chat/completions	Chat completion (streaming)
POST	/v1/messages	Messages-style completion

Ollama Compatible

Method	Endpoint	Description
GET	/api/tags	List all models
POST	/api/chat	Chat completion
POST	/api/generate	Text generation
POST	/api/show	Model info
GET	/api/ps	Loaded models

Mesh Native

Method	Endpoint	Description
GET	/health	Mesh status
GET	/nodes	All nodes with models
GET	/models	Model list
POST	/chat	Mesh routing
POST	/pull	Download model

Chat Completion Example

OpenAI-compatible request:

curl -X POST http://localhost/v1/chat/completions \
  -H 'Content-Type: application/json' \
  -d '{
    "model": "llama3.1:8b",
    "messages": [{"role": "user", "content": "hello"}],
    "stream": true
  }'

Model Aliases

Use these names to route to the best available model:

• parley:best — Most capable model
• parley:code — Best code model
• parley:fast — Lowest latency
• parley:reason — Best reasoning