> For the complete documentation index, see [llms.txt](https://help.cubelet.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.cubelet.ai/platform-and-integration/connecting-a-client/chatgpt.md).

# ChatGPT

*Connect ChatGPT to a cubelet.ai MCP server via Developer Mode — with an authentication limitation to know before you start.*

## Read this first: the auth limitation

ChatGPT's developer-mode connectors support exactly two authentication modes: an OAuth sign-in flow, or no authentication at all. There is currently no way to attach a static API key, bearer token, or custom request header from the Connectors UI, and ChatGPT does not support machine-to-machine grants like client credentials.

The cubelet.ai fleet currently authenticates with a static bearer token (see Auth & access) — a shape ChatGPT connectors can't present. In practice, fleet servers are not reachable from ChatGPT today; a fleet server would have to be fronted by an OAuth sign-in flow first. Use Claude Code or Cursor instead — both accept static headers directly.

## Enable Developer Mode

Settings → Connectors → Advanced → toggle **Developer Mode**. Developer Mode and custom connectors are available on Plus, Pro, Business, Enterprise, and Education plans; the Free plan doesn't support custom connectors.

## Add the connector

1. Open the **Connectors** tab and choose **Add custom connector**.
2. Paste the server's `/mcp` URL — find it on the MCP servers page.
3. Complete the OAuth sign-in if the server prompts for one.
4. Enable the connector for the conversations where you want ChatGPT to use it.

ChatGPT only connects to remote MCP servers over HTTPS; it doesn't support local/stdio servers. That isn't a limitation here, since every cubelet.ai fleet server is a remote HTTPS endpoint.