MCP Catalogs
Homefieldflow screenshot

fieldflow

by guillaumegay13·111·Score 47

FieldFlow converts OpenAPI endpoints into selectively filtered tools with MCP support for AI clients like Claude Desktop.

developer-toolsai-llmproductivity
11
Forks
2
Open issues
this month
Last commit
2d ago
Indexed

Overview

FieldFlow is a Python-based tool that transforms OpenAPI-described REST endpoints into selectively filtered tools. It dynamically generates Pydantic models and FastAPI routes that proxy requests to upstream APIs while returning only the fields specified by the caller. The project offers both HTTP server and MCP server modes, with the latter exposing OpenAPI tools through Model Context Protocol. It supports authentication schemes, nested field selection with JSONPath-like syntax, and includes CLI tools for reducing noisy JSON output.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

you:Reducing API response payload size for AI models to minimize context window usage
you:Creating filtered API access for AI applications through MCP protocol
you:Simplifying complex API responses by selecting only relevant fields
you:What authentication methods does FieldFlow support?
you:How does FieldFlow reduce response payload size?

When to choose this

Choose FieldFlow when working with OpenAPI-compliant APIs and need precise field selection to minimize token usage in LLM interactions.

When NOT to choose this

Don't choose FieldFlow if you need write operations or direct database access, as it only provides read-only API proxy functionality.

Tools this server exposes

5 tools extracted from the README
  • get_user_info

    Fetch user information from JSONPlaceholder API with field selection

  • list_posts

    List posts for a user from JSONPlaceholder API with field selection

  • pokemon_read

    Retrieve Pokemon information from PokeAPI with nested field selection

  • ability_list

    List Pokemon abilities from PokeAPI with field selection

  • ability_read

    Read a specific Pokemon ability by ID from PokeAPI

Note: Tools are derived from OpenAPI specs and exposed through MCP server. Example tool calls shown in README demonstrate the field selection functionality.

Comparable tools

openapi-mcpapi-mcpfastmcpcurl-mcp

Installation

# Install with MCP support
pip install -e '.[mcp]'

# Run HTTP server
fieldflow serve-http

# Run MCP server (for Claude Desktop)
fieldflow-mcp

For Claude Desktop integration, add to config:

{
  "mcpServers": {
    "fieldflow": {
      "command": "python",
      "args": ["-m", "fieldflow_mcp.cli"]
    }
  }
}

FAQ

What authentication methods does FieldFlow support?
FieldFlow supports Bearer tokens (OAuth 2.0, JWT), API keys, Basic authentication, and automatic handling of OpenAPI-defined security schemes.
How does FieldFlow reduce response payload size?
FieldFlow accepts a 'fields' parameter in requests that allows selecting specific data fields. It uses JSONPath-like syntax to select nested fields and arrays, returning only the requested data.

On Hacker News

Recent discussion from the developer community.

Compare fieldflow with

GitHub →

Last updated · Auto-generated from public README + GitHub signals.