MCP integrations
Set up PrompTessor MCP
Setup and integration guides for ChatGPT, Claude, Codex, Claude Code, Gemini CLI, Cursor, VS Code, Devin, and other MCP clients.
Verified against the implementation ·
Prerequisites
- 1
Use an eligible PrompTessor account
You need a verified PrompTessor account and an active plan that includes Public API and MCP access.
- 2
Use the server URL, not an API key
All integrations on this page connect to the hosted remote MCP server through OAuth. Sign in to PrompTessor when the client opens your browser; there is no key to copy or paste.
- 3
Use a compatible client
The client must support remote Streamable HTTP servers and MCP OAuth discovery. Use the current client version when possible.
- 4
Allow browser sign-in
OAuth requires a browser and, for local CLI clients, access to the localhost callback opened during authorization.
- 5
Check workspace policy
Managed ChatGPT, Claude, Cursor, VS Code, or Devin workspaces can require an administrator to enable custom MCP connections.
- Name
PrompTessor- Server URL
https://mcp.promptessor.com/mcp- Transport
Streamable HTTP- Authentication
OAuth- API key
Not required and not accepted for MCP
- Use the exact HTTPS URL shown above, including the /mcp path.
- Choose HTTP or Streamable HTTP. Do not choose the legacy SSE transport.
- Leave custom client ID, client secret, API key, and headers empty. PrompTessor advertises its OAuth configuration automatically.
- Review the requested scopes before approving access. Tool, prompt, and resource support can vary by client.
Safe first test
After connecting, ask the client to call get_usage. It confirms the connection, account, scopes, and usage limits without consuming an AI workflow credit.
Choose an MCP client
Open the focused setup guide for your client. Every guide uses the same hosted PrompTessor MCP server and OAuth authorization flow.