🦎Psyllama

Integrations

Psyllama integrates with any tool that can talk HTTP. The server exposes an OpenAI-compatible API, so many existing clients can be pointed at http://localhost:11434.

Common integration patterns

- Chat applications and desktop clients
- RAG pipelines (embeddings + vector DB)
- IDE assistants
- Automation / agents using tool-calling

Using OpenAI-compatible clients

Configure the base URL to Psyllama and use a local-only key if the client requires one.

Using the native API

See the API Reference section for endpoints.

curl http://localhost:11434/api/chat \
  -d '{"model":"kimi-k2.5:cloud","messages":[{"role":"user","content":"Hello"}]}'