@cnoe-io
基本情報
OpenAPI to MCP Server Code Generator
以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。
{
"mcpServers": {
"openapi-mcp-codegen": {
"command": "uvx",
"args": [
"--from",
"git+https://github.com/cnoe-io/openapi-mcp-codegen.git",
"openapi_mcp_codegen",
"\\"
]
}
}
}ツール一覧はまだありません
この掲載には表示できるツール一覧がまだありません。サーバーにツールがないという意味ではありません。対応機能はプロジェクトのドキュメントをご確認ください。
上のボタンで一時的なサンドボックスでこのサーバーを実行し、ツール一覧を取得できます。
This tool generates Model Context Protocol (MCP) servers from OpenAPI specifications, producing Python packages that AI assistants can use to interact with APIs. It parses paths, operations, and schemas from OpenAPI specs (JSON or YAML) and renders Jinja2 templates into a structured MCP server with tools, models, and client code.
Run uvx --from git+https://github.com/cnoe-io/openapi-mcp-codegen.git openapi_mcp_codegen --spec-file <path> --output-dir <path> with optional flags such as --generate-agent, --generate-eval, or --enable-slim. For enhanced documentation, use python -m openapi_mcp_codegen.enhance_and_generate.
A Python package with FastMCP-based server, tool functions for each API operation (async, type-hinted), an API client module, configuration files (pyproject.toml, .env.example), and optional agent/evaluation components.
Python 3.8+, `uv` package manager, and the Python packages `mcp>=1.9.0`, `httpx`, `Jinja2`, and `Ruff`. The generated server runs in STDIO mode by default.
The README does not specify version limits; it processes OpenAPI specs in JSON or YAML format. The generator uses standard path, operation, and schema parsing.
By default the MCP server communicates over STDIO using the Model Context Protocol. When combined with `--generate-agent`, an A2A server is also produced for HTTP-based agent communication. SLIM transport is available experimentally.
Yes, use the `--dry-run` flag to preview generation output without creating files.
Automated codebase documentation for GitHub: searchable docs, MCP context for AI agents, doc drift detection, and Friday Cleanup PRs.
One case file per task — shared by every session, agent and person that touches it. Self-hosted MCP server and live board. MIT.
Verified distributed compute for humans, applications, and agents. Submit Blender, Octane, and video workloads, route them across qualified GPUs, track asynchronous execution, download finished artifacts, and verify FARP
Interactive MCP App demo built on Metabind: personal-finance answers render as UI cards (net worth, spending, subscriptions) in Claude and ChatGPT, and natively in iOS/Android apps.
Gives AI agents public URLs (tunnels) for localhost, live HTTP traffic inspection, snapshot publishing, and access control. Part of LocalCan, the ngrok alternative for Mac, Windows and Linux. Free plan.
コメント