Open WebUI¶

Open WebUI connects to any OpenAI-compatible backend through its Connections settings — Modelship qualifies without any special support on either side.

In Open WebUI, go to Settings → Admin Settings → Connections → OpenAI API.
Set the API Base URL to http://<modelship-host>:8000/v1.
Set the API Key to a real key if MSHIP_API_KEYS is configured, or any non-empty placeholder if not.
Save — Open WebUI queries GET /v1/models to populate the model picker with whatever is in your models.yaml.

Or via environment variables if you're running Open WebUI in Docker:

docker run -d \
  -e OPENAI_API_BASE_URL=http://<modelship-host>:8000/v1 \
  -e OPENAI_API_KEY=your-key-or-placeholder \
  -p 3000:8080 \
  ghcr.io/open-webui/open-webui:main

Chat, streaming, and image generation all work through the standard /v1/chat/completions and /v1/images/generations endpoints. Which higher-level Open WebUI features (e.g. its own tool-calling UI) light up depends on Open WebUI's own client-side detection of what the connected backend supports — that behavior is Open WebUI's, not something Modelship controls or has specifically verified feature-by-feature.