> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.joytokens.ai/llms.txt.

# Claude Desktop

> Use Claude Desktop with JoyToken

## Enable Third-Party Inference

1. In Claude Desktop, open **Help > Troubleshooting** and enable Developer Mode.
2. Open **Developer > Configure Third-Party Inference**.
3. Select **Gateway** as the inference provider.

## Configure JoyToken

| Setting             | Value                                |
| ------------------- | ------------------------------------ |
| Gateway base URL    | `https://api.joytokens.ai/anthropic` |
| Gateway API key     | Your JoyToken API key                |
| Credential kind     | Static API key                       |
| Gateway auth scheme | `x-api-key`                          |
| Model discovery     | Disabled                             |
| Inference model     | `auto`                               |

Apply the configuration locally, then start a new chat.

## Verify

Send `Reply with exactly: pong`. A successful response is `pong`.

## Common Errors

| Error                         | Fix                                                   |
| ----------------------------- | ----------------------------------------------------- |
| `401 Unauthorized`            | Check the Gateway API key and `x-api-key` auth scheme |
| No model is available         | Disable model discovery and add `auto` manually       |
| Request uses another provider | Confirm the inference provider is set to Gateway      |

## Resources

* [Claude Desktop gateway setup](https://claude.com/docs/third-party/claude-desktop/gateway)
* [Claude Desktop in-app configuration](https://claude.com/docs/third-party/claude-desktop/in-app-configuration)
* [JoyToken Environments](/user-guide/environments)