The Claude directory listing is under review. Once it’s approved, the connector becomes discoverable in Claude and the manual steps below become optional. See availability.
ChatGPT
Perform.AI is available as a plugin in ChatGPT. The plugin connects tohttps://mcp.perform.ai/mcp-invoke, the URL its listing shows. See Tools in the ChatGPT plugin for how it differs from https://mcp.perform.ai/mcp.
- In the ChatGPT sidebar, select Plugins.
- Search for
Perform.AI, then select + next to Perform.AI. Once added, it appears under Installed. - Sign in to Perform.AI when prompted. It’s the same sign-in as every other client: your Perform.AI username and password, or Sign in with Google.
- In a conversation, add Perform.AI to your message and ask a question, for example: “Which carriers had the most delayed deliveries to Germany in July 2026?”
- If you belong to more than one organization, the assistant lists them and asks which one to use.
ChatGPT lists Select organization under Write tools because it changes which organization the session acts on. It doesn’t modify your Perform.AI data. Turning it off prevents the assistant from selecting an organization, so leave it on.
Claude
Claude (web, desktop, and mobile):- Go to Settings → Connectors and select Add custom connector.
- Name it
Perform.AIand enter the server URL:https://mcp.perform.ai/mcp. - Select Add, then Connect. The Perform.AI sign-in opens; sign in and approve.
- In a chat, enable the connector from the tools menu and try: “Which carriers do we have configured?”
/mcp inside a session to complete the sign-in.
Microsoft Copilot
A first-party Copilot listing is planned. Meanwhile, MCP-capable Microsoft surfaces can connect manually: Copilot Studio (custom agents): add a new Tool → Model Context Protocol, set the server URL tohttps://mcp.perform.ai/mcp with OAuth authentication, and sign in with your Perform.AI account.
VS Code (GitHub Copilot agent mode): add to your mcp.json:
Clients without remote-server support
Some MCP clients only launch local servers and can’t connect to a remote URL with OAuth directly. For those, the open-sourcemcp-remote bridge runs locally, handles the OAuth flow, and presents Perform.AI to the client as a local server.
Install it once (recommended, so first connection doesn’t stall on a download):
claude_desktop_config.json-style file (macOS: ~/Library/Application Support/Claude/claude_desktop_config.json; Windows: %APPDATA%\Claude\claude_desktop_config.json):
node --version to check).
Custom applications
The server speaks standard MCP over Streamable HTTP with OAuth 2.1, including metadata discovery and dynamic client registration, so a spec-compliant client needs no pre-registered credentials:- Protected resource metadata:
https://mcp.perform.ai/.well-known/oauth-protected-resource/mcp. It names the Perform.AI authorization server. An unauthenticated request returns401with aWWW-Authenticateheader pointing to this document. - Grant types:
authorization_code(with PKCE) andrefresh_token
TypeScript
Redirect URIs: client registrations must use HTTPS redirect URIs (loopback URIs for desktop clients are always accepted). If your client redirects to a domain Perform.AI doesn’t recognize, users see a consent screen naming the destination before the redirect proceeds. To allowlist your redirect domains, contact your Key Account Manager or cs@perform.ai. Claude Desktop and Claude Code users never see this screen.