ICON MCP v109

You can expose ICON MCP v109 endpoints through a dedicated MCP server that lets AI agents and clients interact with ICON MCP v109 using a standardized tool interface. This server handles API access, supports async operations, and can be deployed via Docker or run locally for development and testing.

How to use

Connect to the MCP server from your MCP client or adapter using the provided endpoint. You can discover the available tools, then invoke them to query ICON MCP v109 endpoints and receive structured responses. The server exposes an /mcp path for client connections, and you can run health checks to verify the instance is up.

How to install

Prerequisites you need before starting: Docker (for containerized deployment) or a Python environment with uv and pip for manual setup. You will also use a local MCP client to connect to the server.

# Option A: Using Docker (recommended)
# Start from a clone of the project repository
# (execute in your shell)

# 1) Clone the repository
git clone https://github.com/Traia-IO/icon-mcp-v109-mcp-server.git
cd icon-mcp-v109-mcp-server

# 2) Run locally with Docker (uses the included script)
./run_local_docker.sh

# Option B: Using Docker Compose
# Step 1: Create an environment file
PORT=8000

# Step 2: Start the services
docker-compose up

# Option C: Manual installation (Python + uvx)
# Step 1: Install dependencies
uv pip install -e .

# Step 2: Run the server
uv run python -m server

Configuration and environment

Environment variables you can configure to tailor the server behavior include port, deployment stage, and log level.

PORT=8000
STAGE=MAINNET
LOG_LEVEL=INFO

Health check and basic usage

After starting the server, perform a health check to confirm it is running correctly. Use your preferred method to invoke the health endpoint or a provided health check script.

# Example health check command (adjust as needed)
python mcp_health_check.py

Using with CrewAI

You can connect your CrewAI workflow to the MCP server to list and use tools exposed by ICON MCP v109. Create an adapter and point it at the server URL, then iterate over available tools and execute a selected tool in your workflow.

from traia_iatp.mcp.traia_mcp_adapter import create_mcp_adapter

with create_mcp_adapter(
    url="http://localhost:8000/mcp/"
) as tools:
    for tool in tools:
        print(f"Available tool: {tool.name}")
    # Example usage
    result = await tool.example_tool(query="test")
    print(result)

Available tools

example_tool

Placeholder tool (to be implemented)