🤖 Solana Trading MCP Server

🖥️ Live Demo → An AI-powered Solana trading bot with a live dashboard, built on the MCP (Model Context Protocol) framework.

Features

• MCP Server — exposes Solana trading tools via HTTP for any AI agent

• AI Agent — connects a local Ollama model (hermes3) to control the bot

• Live Dashboard — monitor trades, balances, and bot status

• Bot Studio — configure and control the bot from the browser

• Transaction Trail — full history of all transactions

• Paper Mode — test without real money

Related MCP server: @blockrun/alpha

Tech Stack

• Python, FastAPI, FastMCP 2.0

• Solana, solders, anchorpy

• Streamlit + Plotly dashboard

• LangChain + Ollama (hermes3)

• SQLite + Redis for state

Setup

1. Clone the repo

git clone https://github.com/itsKazgar/solana-mcp-server.git
cd solana-mcp-server

2. Create virtual environment

python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

3. Copy and fill in your config

cp .env.example .env

4. Replace YOUR_IP in agent.py and mcp_config.json with your server IP

5. Start the server

python main.py

6. Start the dashboard

streamlit run dashboard/app.py

Security

• Never commit your .env file

• Use PAPER_MODE=true before trading real funds

• Generate a strong token: openssl rand -hex 32

License

MIT