Open WebUI

Moth DocsTutorials

A chat app you host yourself, which lists every Moth chat model once it has one connection.

In the app

  1. Add a connection

    Open Admin Settings, then Connections, and press the plus beside the OpenAI API connections.

  2. Fill it in

    URL https://api.runmoth.si/v1 and API Key your moth_ key. Leave Model IDs empty to list every model, or add IDs there to show only those.

  3. Check it worked

    Press Verify Connection, then Save. Moth's model IDs appear in the model picker, and a test chat streams its reply.

Or when you start it

The same connection can be set when the container starts. Settings saved in the app win over these once it has run, so change them in the app afterwards.

shell
docker run -d -p 3000:8080 \
  -v open-webui:/app/backend/data \
  -e WEBUI_SECRET_KEY="$(openssl rand -hex 32)" \
  -e OPENAI_API_BASE_URL=https://api.runmoth.si/v1 \
  -e OPENAI_API_KEY=moth_your_key_here \
  -e TASK_MODEL_EXTERNAL=deepseek-v4.1-flash \
  -e ENABLE_OLLAMA_API=False \
  --name open-webui --restart always \
  ghcr.io/open-webui/open-webui:main

Good to know

  • Open WebUI uses a model for chat titles and other small tasks, and each is a paid call, so TASK_MODEL_EXTERNAL points them at a low-priced model.
  • Document upload keeps working, because Open WebUI embeds documents on your own machine by default. Don't switch its embedding engine to OpenAI with Moth's address, since Moth doesn't serve embeddings.
  • Leave image generation off, since Moth's images run as jobs rather than through OpenAI's images endpoint.
  • Give the app a key of its own with a spend cap, especially if other people use it.