This Zerodha MCP Server provides a Python-based backend that connects to Zerodha's trading platform and exposes trading operations via the Model Context Protocol (MCP). It supports account management, placing and modifying orders, retrieving holdings, margins, positions, and order histories, as well as generating login URLs and access tokens through the MCP-enabled endpoints. The server can be interacted with through multiple clients (Agno-based, Google ADK-based, or Gradio) to provide an interactive, natural-language interface for trading operations. To start, ensure your environment variables contain your Zerodha API credentials and OpenAI/Google keys as needed. Once running, you can use the built-in tools like get_login_url, get_access_token, get_user_profile, get_margins, get_holdings, get_positions, get_orders, get_order_history, get_order_trades, place_order, modify_order, and cancel_order to manage your Zerodha account via MCP-compatible clients. The included clients (Agno, Gradio, and Google ADK) connect to the MCP server and present an accessible UI or chat-based interface for executing these actions.

To interact using the server directly, connect a client to the MCP host/port you started the server on, then issue commands through the MCP interface. The Agno client provides a conversational interface, the Gradio client offers a web UI, and the Google ADK client enables integration with Google ADK workflows. Each client authenticates with the MCP server and forwards user intents to Zerodha through the server.

Prerequisites:

• Python 3.8+ installed
• Git installed
• Zerodha account with Personal API access and API key/secret
• Optional: OpenAI API key (for Agno/AI-assisted chat) and Google API key (for Google ADK)

Install steps:

1. Clone the repository: git clone https://github.com/mtwn105/zerodha-mcp-server-client.git cd zerodha-mcp-server-client

2. Create and activate a virtual environment: python -m venv venv source venv/bin/activate # On Windows: venv\Scripts\activate

3. Install dependencies: pip install -r requirements.txt

4. Configure environment variables:

   • Copy the example env file and edit values cp .env.example .env

     Edit .env with your credentials and settings

5. Run the server (example): python server.py

6. Optional: Run a client (Agno, Gradio, or Google ADK) to connect to the MCP server using host/port from the .env file or CLI arguments.

Tips and notes:

• Ensure ZERODHA_API_KEY and ZERODHA_API_SECRET are kept secret and not committed to version control.
• The SERVER_MODE can be 'sse' (Server-Sent Events) or 'stdio' depending on your client integration; adjust in .env accordingly.
• If you encounter authentication issues, re-authenticate with Zerodha and refresh tokens as required by Zerodha's API workflow.
• The MCP_HOST and MCP_PORT settings determine where clients connect; keep them consistent across server and clients.
• The Gradio client requires OPENAI_API_KEY for AI-assisted features; ensure the key is valid and has access.
• For Google ADK integration, ensure GOOGLE_API_KEY or appropriate Application Default Credentials are configured.
• Review the tools list (get_login_url, get_access_token, get_user_profile, get_margins, get_holdings, get_positions, get_orders, get_order_history, get_order_trades, place_order, modify_order, cancel_order) to plan automation flows and natural language commands.
• If running multiple MCP servers, use distinct host/port pairs and environment configurations to avoid conflicts.