Provides an HTTP MCP endpoint and a local CLI runtime to interact with the IB REST API via MCP.
Configuration
View docs{
"mcpServers": {
"rcontesti-ib_mcp": {
"url": "http://localhost:5002/mcp/",
"headers": {
"MCP_SERVER_HOST": "0.0.0.0",
"MCP_SERVER_PATH": "/mcp",
"MCP_SERVER_PORT": "5002",
"MCP_TRANSPORT_PROTOCOL": "streamable-http"
}
}
}
}This MCP server provides a Model Context Protocol interface on top of Interactive Brokers REST API, using the Web API as the data source. It exposes a fast, scriptable interface for querying IB data and issuing standard actions through a simple HTTP endpoint and a local, self-contained MCP runtime.
You connect to the MCP server via an HTTP client or through a local, command-based interface. Use the HTTP endpoint to query IB data and perform standard actions in a RESTful, JSON-friendly way. If you prefer a local runtime integration, run the MCP server as a stdio process and communicate with it using the provided command-line entry points.
Prerequisites you need on your machine are Docker Desktop and Git, along with a modern Python runtime for the MCP server if you opt to run it directly. Follow these steps to set up the containerized MCP server stack.
# Clone the repository
git clone https://github.com/rcontesti/IB_MCP.git
# Navigate to the project directory
cd IB_MCP
# Copy the .env.example file to .env and edit as needed
cp .env.example .env
# Build and start the images (detached)
docker compose up --build -dAfter the images are up, authenticate through the Client Portal Gateway UI at a URL like https://localhost:5055/ to establish a session. This MCP stack uses a gateway with a tickler service to keep sessions alive. The MCP server itself is exposed for HTTP access at http://localhost:5002/mcp/.
Add the MCP server connection in your editor by configuring the MCP server entry. Use the values shown to point your editor to the local MCP endpoint.