Install Open WebUI first
For a local Open WebUI test, install Docker Desktop, start the official Open WebUI container, open the UI on localhost, and then add CorvusLLM as an OpenAI-compatible connection.
| What to install | Official source | Quick check |
|---|---|---|
| Docker Desktop | Docker Desktop docs | docker --version |
| Open WebUI | Open WebUI quick start | http://localhost:3000 after the container starts. |
| CorvusLLM key | Buy or top up balance | Paste as the OpenAI-compatible API key. |
docker --version
docker run -d -p 3000:8080 -v open-webui:/app/backend/data --name open-webui ghcr.io/open-webui/open-webui:main Zero-to-working setup
Use this block when you want the shortest reliable path from a clean install to one successful Open WebUI request through CorvusLLM.
- Docker Desktop or another supported Open WebUI install path
- Open WebUI quick start instance
- CorvusLLM key
Official source: Docker Desktop docs, Open WebUI quick start
Open Open WebUI, then use Admin Panel > Settings > Connections and add or edit an OpenAI-compatible connection.
- Open Admin Panel > Settings > Connections
- Add or edit an OpenAI-compatible connection
- Paste https://base.corvusllm.com/v1 and your CorvusLLM key
- Refresh or cache the base model list if your build requires it
- Start a chat with gpt-5.5 and send Say open-webui-ok
A new Open WebUI chat with gpt-5.5 returns the short test answer through the saved connection.
| Value | Paste this |
|---|---|
| Connection type | OpenAI-compatible or custom OpenAI |
| Base URL | https://base.corvusllm.com/v1 |
| API key | YOUR_CORVUSLLM_KEY |
| Starter model | gpt-5.5 |
Do not paste /chat/completions into Open WebUI base URL fields. Use only the /v1 base URL.
Connection values
Open Open WebUI in your browser, then go to Admin Panel, Settings, and Connections. Add or edit an OpenAI-compatible connection and paste these values.
| Open WebUI field | Where to change it | Value to paste |
|---|---|---|
| Connection type | Admin Panel > Settings > Connections > OpenAI-compatible connection | OpenAI-compatible or custom OpenAI |
| Base URL | The connection's API/Base URL input | https://base.corvusllm.com/v1 |
| API key | The same connection's API key input | Your delivered CorvusLLM key |
| First model | Model picker after the connection is saved | gpt-5.5 |
Look for the OpenAI-compatible provider or connection screen. The important parts are the /v1 base URL, your CorvusLLM key, and a supported model slug.
Enable Settings > Connections > Cache Base Model List and keep only the model slugs you actually use in the Open WebUI model list.
Connection test
After saving the OpenAI-compatible connection, start a new chat with the starter slug and send a short message such as Say open-webui-ok. This confirms that Open WebUI is using the CorvusLLM key, the /v1 base URL, and a valid public model slug.
- If the model list does not appear, refresh the connection or paste the model slug manually.
- If the chat returns an auth error, recopy the key from the dashboard.
- If a single model fails, try another public slug and check service status.
After it connects
After the first successful chat, you can switch to other supported slugs from Models & Slugs, including Claude and GLM families through the same OpenAI-compatible path. Treat the live catalog as the source of truth for any additional families.
Recommended live catalog picks
This block follows the current public catalog, so new models and reordered families can surface here automatically after you save them.
Loading current catalog
CorvusLLM is pulling the latest recommended models for Open WebUI.