Integrations¶
Modelship speaks the OpenAI API (/v1/chat/completions, /v1/responses,
/v1/embeddings, /v1/audio/*, /v1/images/generations), so anything that
already talks to OpenAI — an SDK, a chat UI, a low-code agent builder — can
point at Modelship instead by changing its base URL.
All of these guides assume a running Modelship gateway reachable at
http://<host>:8000 (see Quick start if you don't
have one yet). Modelship doesn't check the API key, so any non-empty string
works, unless an auth layer in front of it expects a real one; most
OpenAI-compatible clients require some value in that field.
- OpenAI SDK — Python and JS/TS, the direct integration path
- Open WebUI — a self-hosted chat UI
- Dify — a low-code agent/app builder
- n8n — workflow automation
- LiteLLM — an API-key, budget and rate-limit layer in front of Modelship
- Responses-speaking agents — agent frameworks and
custom loops built on
/v1/responsesspecifically