Skip to content

ChatGPT

Time: about 5 minutes · You need: ParleML reachable on public HTTPS, and a ChatGPT plan that allows custom connectors (developer mode, or a company workspace). You'll end with: ChatGPT answering questions about your ParleML workspaces.

ChatGPT reaches ParleML as a connector. It calls /mcp from OpenAI's servers, so it cannot reach a ParleML on localhost (see which clients can reach a local ParleML). There is no token to paste: ChatGPT signs in through OAuth.

1. Add the connector

In ChatGPT, add a custom connector with your ParleML MCP address, such as https://parleml.example.com/mcp. The address is under Settings → Access → Your own agent in ParleML.

2. Approve the access

ChatGPT opens ParleML's consent page in your browser.

  1. Sign in to ParleML if asked.
  2. Choose what to allow: read; read and check (compute); or, if you are a ParleML administrator, build in the workspaces you pick, within a run budget.
  3. Approve.

3. Ask something

In a ChatGPT conversation with the connector on, ask What workspaces do I have in ParleML?

The connection shows under Settings → Access → Connected agents in ParleML. Choose Disconnect there to end it at any time.

What works

Every tool of the scope you allowed: the read tools (catalog, SQL, profile, EDA, evaluation, insights, charts, with PNG charts where ChatGPT displays images), the compute tools, and the builder tools (Tokens & scopes).

Details: how the sign-in works

ParleML is its own OAuth 2.1 authorization server. ChatGPT discovers it from the /mcp endpoint, registers itself and opens the consent page. If ParleML sits behind a proxy that rewrites the host, set PARLEML_OAUTH_ISSUER to ParleML's public origin so the discovery URLs are right.

Next