Convert a Postman collection to MCP

Import a collection

Export a Postman v2.x collection as JSON and import it through mcpmint. Requests and folders become an operation catalog you can inspect and select. Request URLs, methods, parameters, bodies and supported authentication are normalized into the same model used for OpenAPI imports.

A collection may contain concrete examples rather than complete schemas. Postman scripts, environment execution and pre-request logic are not a Postman runtime inside the generated MCP server. Review inferred inputs, unresolved variables and validation warnings, and implement any necessary behavior in the exported source.

Prepare the collection safely

Remove tokens, cookies and passwords before importing or sharing a collection. Replace private values with placeholders and configure the upstream base URL and authentication in the generated server's environment.

1. Export the collection as JSON, keeping it under the 5 MiB specification limit.
2. Import the file and inspect the capabilities and endpoint warnings.
3. Rename tools and clarify descriptions so an AI client can choose the right operation.
4. Select no more than 500 tools, preview the code and run local request mocks.
5. Export TypeScript or Python, review the README and test against a development API before connecting an AI client.

Understand live tests

Inspection and mocks stay in your browser. Run live requests locally using the CLI or generated server against a development API. The CLI requires explicit permission for mutations. Inspect the outgoing request and review authentication before testing.