Show HN: Transform your REST API into optimized MCP server in one command line (github.com)

FieldFlow is an open‑source tool that converts OpenAPI 3.0 specs into ready-to-run, selectively filtered tool endpoints in one command. It dynamically generates Pydantic request/response models (preserving aliases and optional fields), builds FastAPI routes that accept parameters plus an optional "fields" list to slice responses, and proxies upstream requests via httpx while formatting path and query parameters. It works with any OpenAPI-compliant JSON/YAML (including nested schemas and $ref), exposes automatic docs at /docs, and includes a CLI and example specs (jsonplaceholder, PokeAPI) for quick testing. The project also provides an optional MCP (Model Context Protocol) layer so the same generated tools can be surfaced to LLM clients like Claude Desktop—FieldFlow can run an MCP server over stdio and Claude will automatically discover and invoke the tools. For the AI/ML community this matters because it removes boilerplate for turning REST APIs into LLM-friendly tools, reduces overfetching by returning only requested fields (improving latency, bandwidth and privacy), and enables seamless local tool integration for agent workflows. Configuration is environment-driven (FIELD_FLOW_OPENAPI_SPEC_PATH / FIELD_FLOW_TARGET_API_BASE_URL), it’s MIT‑licensed, and includes tests and contributor docs for extensibility.