Skip to main content

MCP Integration Guide

Overview​

This guide will help you get started with the Odos MCP Client—a Python-based Model Context Protocol server exposing Odos DEX aggregation and Zerion portfolio analytics tools. You’ll learn to:

Install & configure the MCP client
Run the MCP server
Integrate the MCP server into Cursor

For detailed reference on every tool, see the odos-mcp README.

Setup​

Make sure you have:

  • Python 3.10+
  • poetry (or pip)

Step 1: Clone & Install​

Clone the repository

git clone [email protected]:odos-xyz/odos-mcp.git
cd odos-mcp

Create & activate a virtual environment (Optional)

python -m venv .venv
source .venv/bin/activate

Install dependencies​

poetry install

# Alternatively, if you use pip:

pip install -r requirements.txt

Step 2: Portfolio Analytics Configuration (Optional)​

Create a .env file in the project root with the following command:

dotenv

Add Zerion API key to your .env​

ZERION_API_BASE="https://api.zerion.io"

WARNING: Never commit your .env or private key to source control.

Step 3: Start the MCP Server​

poetry run python -m src.client.server

The MCP server will appear to hang because it streams stdio by default and does not output console messages.

Step 4: Integrate with Cursor​

Cursor supports both stdio and Streamable HTTP. We'll configure Cursor to access the MCP over Streamable HTTP for this tutorial.

Start Odos MCP in Streamable HTTP mode​

poetry run python -c 'from client.server import mcp; mcp.run(transport="streamable-http")'

You will see a confirmation message that the web server is running.

INFO:     Uvicorn running on http://127.0.0.1:8000 (Press CTRL+C to quit)

Cursor UI Setup​

Open Cursor Settings > Tools & Integrations.

Click New MCP Server

Enter the following in your mcp.json file and click Save:

{
  "mcpServers": {
    "odos-mcp": {
      "transport": "streamable_http",
      "url": "http://localhost:8000/mcp",
      "env": {
        "ZERION_API_BASE":  "${ZERION_API_BASE}"
      }
    }
  }
}

Expand odos-mcp to see the available tools.

mcp tools

Hover your mouse over each tool to see a brief description. Full details are available on GitHub.

You may now begin to interact with the MCP in the Cursor agent panel.

use case example

use case example

Ready to Start Building?​