AionMCP - Autonomous Go

Autonomous MCP server that imports OpenAPI, GraphQL, and AsyncAPI specs and exposes them as agent tools.

go

1

GitHub Stars

go

Language

2 months ago

First Indexed

3 weeks ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "kiransth77-aionmcp": {
      "command": "./bin/aionmcp",
      "args": [],
      "env": {
        "AIONMCP_LOG_LEVEL": "debug"
      }
    }
  }
}

AionMCP is an autonomous Go-based MCP server that dynamically imports OpenAPI, GraphQL, and AsyncAPI specifications and exposes them as tools for agents. It supports self-learning, context-aware operation, and auto-generated documentation, all within a clean, maintainable architecture that makes it easy to extend and adapt.

How to use

You run the MCP server locally and interact with it using an MCP client. Start by launching the server, then discover available tools, execute a tool to perform an action, and view learning insights and auto-generated documentation as the system adapts to your usage patterns.

Typical usage workflow:

How to install

Prerequisites and initial setup ensure you have the necessary tooling to build and run the MCP server.

# Prerequisites
# - Go 1.21 or higher
# - Git

# From source
git clone https://github.com/kiransth77/aionmcp.git
cd aionmcp

goflags=''

go mod download

go build -o bin/aionmcp cmd/server/main.go

# Run the server with default configuration
./bin/aionmcp

Configuration and run options

You can customize how the server runs by providing a configuration file or adjusting environment variables at startup.

Example startup with a custom config and debug logging:

# Start with a specific config file
./bin/aionmcp --config config.yaml

# Enable verbose debugging output
AIONMCP_LOG_LEVEL=debug ./bin/aionmcp

Deployment and related usage notes

If you prefer containerized deployment, you can build the binary and run it in a container as a baseline. You can also run the server embedded in a Go application if you want to integrate MCP capabilities directly into your Go project.

Container example (local port mapping):

# Build and run a container image locally
docker build -t aionmcp:latest .
docker run -p 8080:8080 aionmcp:latest

Notes on tools, endpoints, and developer tips

Tools and endpoints are exposed by the server and can be discovered via the MCP client. The server exposes its capabilities through a registry, learning engine, and auto-documentation components, all designed to be dynamically updatable without restarting the service.