Use with Open WebUI
Open WebUI is an open-source chat interface compatible with OpenAI-compatible APIs. You self-host Open WebUI and point it at your CaseDesk account endpoint; model inference remains at the approved customer-controlled or provider-controlled runtime, while the UI runs on your own machine or server.
:::info CaseDesk does not host Open WebUI for customers Each team runs its own Open WebUI instance. The instructions below show how to connect it to your approved CaseDesk production connection. :::
Option 1 — Run Open WebUI locally (Docker)
The simplest way to get started. Runs on your machine in under a minute.
docker run -d \
-p 3000:8080 \
-e OPENAI_API_KEY=cd_live_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx \
-e OPENAI_API_BASE_URL=https://getcasedesk.com/v1 \
-v open-webui:/app/backend/data \
--name open-webui \
--restart always \
ghcr.io/open-webui/open-webui:main
Replace cd_live_xxx... with the account-level production API key shown in CaseDesk Settings. The requested model must be approved on a healthy customer-controlled connection.
Then open http://localhost:3000. Models approved for your connected runtime will appear in the model selector.
Option 2 — Connect an existing Open WebUI installation
If your team already runs a shared Open WebUI instance, add your CaseDesk account as a connection.
In Open WebUI:
- Go to Settings > Connections
- Under OpenAI API, click the + icon to add a new connection
- Set the Base URL to
https://getcasedesk.com/v1 - Set the API Key to your production API key (
cd_live_...) - Click Save
Models approved for your active connection will appear in the model selector.
Endpoint reference
| Field | Value |
|---|---|
| Base URL | https://getcasedesk.com/v1 |
| API Key | cd_live_xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx |
Connecting other applications
The same endpoint and API key work with any OpenAI-compatible application.
Workflow automation
n8n — use the OpenAI node, set a custom base URL, and point it at your CaseDesk endpoint.
Flowise and Langflow support custom OpenAI endpoints directly in their node configuration.
Developer tools
| Tool | Use case |
|---|---|
| Continue.dev | IDE coding assistant |
| Cline | IDE agent |
| LibreChat | Chat interface with RAG |
| Flowise | Visual LLM flow builder |
| Langflow | Visual LLM flow builder |
| n8n AI nodes | Workflow automation |
| OpenAI Python / JS SDK | Custom applications |
Any tool that accepts a custom base_url and api_key can use CaseDesk once a supported model is available on an approved connection.
Working with real-time information
Open-source models have a training cutoff — they do not know about events after their training data ends. If you have web search configured in your Open WebUI installation (for example via SearXNG), enable it in the chat toolbar to let the model fetch current information before answering.
For specific URLs, use the file/URL attachment feature to paste the page content directly into the conversation context.