Skip to content
RunAPI Developer Docs
EN 中文
Developer Resources
Developer Resources

Open WebUI

Install Open WebUI, add RunAPI as an OpenAI-compatible connection, select a model, and verify a chat.

View Markdown

Connect Open WebUI to RunAPI through its OpenAI-compatible Connection. You need Open WebUI administrator access and a standard RunAPI API key from the Authentication Guide.

Install Open WebUI

Follow the official Open WebUI quick start for your platform. For a local Docker installation, the official image can be started with:

SHELL 
docker run -d -p 3000:8080 -v open-webui:/app/backend/data --name open-webui ghcr.io/open-webui/open-webui:main

Open the application, create the administrator account, and sign in before adding the connection.

Add the RunAPI connection

  1. Open Admin Settings → Connections → OpenAI, then select Add Connection.
  2. Set the URL to https://runapi.ai/v1 and enter your RunAPI API key.
  3. Leave Model IDs (Filter) empty for the first connection attempt so Open WebUI can discover the available models.
  4. Save the connection and wait for verification to finish.

Open WebUI sends the connection check from its backend and uses GET /v1/models for discovery. The RunAPI URL is public HTTPS, so a Docker installation does not need host.docker.internal for this connection.

Select a model

Open a new chat and select a RunAPI model from the chat model selector. Use the Model Catalog to confirm the current model identifier and capabilities.

If discovery does not populate the selector, edit the connection and add the exact identifier under Model IDs (Filter). A model discovery error and a Chat Completions error are separate signals; test the manually added model before removing the connection.

Verify the connection

Send a short prompt such as Reply with the word connected. A normal streamed or complete response confirms that Open WebUI can use the selected RunAPI model.

Troubleshoot

  • Verification returns 401 or 403: replace the API key in the connection with a valid standard key from the Authentication Guide.
  • No RunAPI model appears: confirm the URL ends in /v1, then add one current identifier to Model IDs (Filter) and retry the chat.
  • The model is visible but chat fails: confirm the identifier supports OpenAI-compatible Chat Completions in the Model Catalog, then inspect the Open WebUI backend logs for the request status.
  • A local URL works in the browser but not Open WebUI: Open WebUI connections originate from its backend. This does not affect the public https://runapi.ai/v1 URL, but it matters for other local connections.

For the shared protocol behavior, use the LLM API Quickstart. For request and response fields, use the Chat Completions API Reference.