knowsuchagency/mcp2cli is a Python tool that builds a command-line interface (CLI) at runtime, without code generation, from MCP servers, OpenAPI specs, and GraphQL endpoints. Because the commands are created on the spot from nothing more than an API schema, you don't have to write or build separate client code each time you call it. The release tags and versions published in this repository are not given in the source material.
MCP (Model Context Protocol) is a protocol that connects AI agents with external tools and data, OpenAPI is a specification format that describes the structure of REST APIs, and GraphQL is a query style based on a schema. mcp2cli treats all three the same way. It connects to MCP servers over HTTP/SSE (Server-Sent Events) and stdio (standard input/output) and turns them into CLI commands, and it also supports OAuth authentication, caching, and JSON output. It additionally provides an installable Skill for AI coding agents. The center of this tool is that instead of writing dedicated wrapper code yourself, you can call it straight from the shell as long as you have the schema. It is released under the MIT license, and the repository was created on March 9, 2026. The star count is 2421.
That said, there are no confirmed performance figures or benchmarks. Novelty, technical completeness, and quality scores are absent from the material, and the only score left is a momentum score of 2.1. On a 10-point scale, that is on the low side. How stable this approach is on real APIs with complex responses, and how it handles failed calls, is hard to judge from the material alone. It seems perfectly fine as something to try out in the shell. There is not yet enough evidence to entrust it with the foundation of automation overall.