Algorand agent MCP server

This project implements an MCP server for Algorand agent interactions, available in both TypeScript and Python implementations.

TypeScript Implementation

To experiment with the TypeScript implementation:

1. Run npm run build to generate the output file

2. Create a .env file with the private key and app id

3. Add the MCP server to your AI setup, e.g. to prototype it manually add it to Claude Desktop with a config like this:

   {
     "mcpServers": {
       "algorand": {
         "command": "node",
         "args": ["path/to/algorand-agent-mcp-server/build/index.js"]
       }
     }
   }

Python Implementation

To use the Python implementation:

1. Set up a Python environment (Python 3.8+ recommended)

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

3. Create a .env file with the required parameters (use .env.template as a guide)

4. Run the server: python app.py

5. Add the MCP server to your AI setup, e.g. to prototype it manually add it to Claude Desktop with a config like this:

   {
     "mcpServers": {
       "algorand": {
         "command": "python",
         "args": ["path/to/algorand-agent-mcp-server/app.py"]
       }
     }
   }