Use Claude Desktop with the Gateway¶
Claude Desktop is Anthropic's desktop app for macOS and Windows. Its third-party inference mode can send requests to a gateway instead of Anthropic. Point it at the Gateway to use Claude models hosted in your region.
The Gateway speaks the Anthropic Messages API, which is what Claude Desktop uses in this mode.
Prerequisites¶
- Claude Desktop installed (download)
- An LLM API key and endpoint from my.amazee.io
Connect Claude Desktop to the Gateway¶
- In Claude Desktop, go to Help → Troubleshooting → Enable Developer Mode
- Open Developer → Configure Third-Party Inference…
- Under Connection, set Inference provider to Gateway
- Fill in the gateway settings:
| Field | Value |
|---|---|
| Gateway base URL | https://llm.[your-region].amazee.ai (no /v1) |
| Gateway API key | Your LLM API key |
| Credential kind | The API key option |
| Gateway auth scheme | Bearer |
Replace [your-region] with your region code (e.g. ch103, us103).
- Add the Claude models you want by their Gateway IDs, e.g.
claude-4-5-sonnet. See Available Models. - Save and restart Claude Desktop
Roll out to a team¶
In the same window, click Export to get a .mobileconfig (macOS) or .reg (Windows) file, and deploy it with your device management tool. See Anthropic's gateway deployment guide.
Troubleshooting¶
- Model picker is empty
- Claude Desktop only auto-lists models whose IDs look like Claude models. Add the models by ID in the configuration window.
- "Unauthorized" error
- Check the key is active on the AI keys page in my.amazee.io, and that the base URL uses your key's region.
- Requests still going to Anthropic
- Reopen Developer → Configure Third-Party Inference… and confirm Inference provider is Gateway.