We’re excited to announce the initial release of mcp-graphql-forge (v0.1.0), a lightweight, configuration-driven MCP server that transforms any GraphQL endpoint into an MCP server with zero custom code required.

Released on April 23, 2025, mcp-graphql-forge fills an important gap for teams working with AI agents and GraphQL APIs. Instead of writing custom integration code for every API you want your agents to access, you can now define tools in simple YAML files and let mcp-graphql-forge handle the rest.

What’s New

As the inaugural release, v0.1.0 introduces the complete core functionality of mcp-graphql-forge:

YAML-Driven Tool Configuration
Define your GraphQL queries as tools using clean, human-readable YAML files. Each tool specification includes the query itself, input parameters with type validation, and descriptions that help AI agents understand when and how to use each tool.

Dynamic Tool Discovery
The server automatically discovers and registers all YAML tool definitions in your configuration directory. Add a new tool by dropping in a new YAML file—no code changes or restarts needed.

Flexible GraphQL Integration
Connect to any GraphQL endpoint with configurable authentication via Bearer tokens. The included example demonstrates integration with GitHub’s GraphQL API, but the tool works with any standard GraphQL server.

Multi-Platform Support
Ready-to-use binaries are available for Linux (amd64, arm64, 386), Windows (amd64, 386), and macOS (amd64, arm64). Each release includes checksums for verification and GitHub attestation for provenance.

MCP Protocol Compliance
Serves over the standard stdio protocol, making it compatible with any MCP client implementation including popular AI development environments.

Why It Matters

AI agents are becoming increasingly capable of automating complex workflows, but connecting them to your existing APIs has traditionally required custom integration work. mcp-graphql-forge changes that equation.

For teams already investing in GraphQL APIs, this release means your AI agents can now interact with those APIs through pre-defined, intentional queries without exposing your entire schema. You maintain full control over what data and operations are accessible while eliminating the boilerplate of custom MCP server implementations.

The configuration-driven approach also lowers the barrier to entry. Teams can iterate on tool definitions rapidly without touching code, making it practical to experiment with different agent capabilities and refine them based on real usage patterns.

Getting Started

To get started with mcp-graphql-forge:

  1. Download the pre-built binary for your platform from the GitHub releases page
  2. Configure your GraphQL endpoint and tools using YAML files (see the example/ directory for a complete GitHub API example)
  3. Set the FORGE_CONFIG environment variable to point to your configuration directory
  4. Run the binary to start your MCP server

For developers who prefer building from source, the project requires Go 1.24.2 and can be built with standard Go tooling.

The full documentation includes detailed configuration examples, parameter type specifications, and integration guidance for common MCP clients.

This post was AI-generated using the unsloth/Qwen3.5-122B-A10B-GGUF:Q4_K_M model. For full details on this release, see the v0.1.0 release published on April 23, 2025. Author: release-storyteller