Panther MCP Server

Backtest any trading strategy by describing it to Claude, ChatGPT, or Gemini.

Panther gives your AI assistant the tools to fetch market data, run backtests, and show you the results — no code required. Just describe your strategy in plain English.

Quick Start

1. Get your API key

Sign up free at panther.watch and copy your API key.

2. Install

Claude Desktop — add to your claude_desktop_config.json:

{
  "mcpServers": {
    "panther": {
      "command": "uvx",
      "args": ["panther-mcp"],
      "env": {
        "PANTHER_API_KEY": "pthr_your_key_here"
      }
    }
  }
}

Claude Code:

claude mcp add panther -e PANTHER_API_KEY=pthr_your_key -- uvx panther-mcp

ChatGPT (Pro/Team/Enterprise — Developer Mode):

Add as custom MCP server: https://panther.watch/mcp/

Gemini CLI:

{"mcpServers": {"panther": {"command": "uvx", "args": ["panther-mcp"], "env": {"PANTHER_API_KEY": "pthr_..."}}}}

Any MCP client (Cline, Cursor, VS Code Copilot, etc.):

uvx panther-mcp

3. Start backtesting

Just tell your AI what strategy you want to test:

"Backtest a golden cross strategy (SMA 50/200) on BTC/USDT over 2024"

You'll get results including return, Sharpe ratio, drawdown, every trade, and a shareable link to the full report.

Related MCP server: vibetrader-mcp

Available Tools

Supported Assets

Features

• 25 indicators — SMA, EMA, RSI, MACD, BB, VWAP, ATR, Stochastic, ADX, OBV, Supertrend, Ichimoku, and 13 more

• Short selling — Long, short, or bidirectional strategies

• Strategy optimization — Parameter sweeps with constraints and ranking by any metric

• Portfolio backtesting — Test across multiple weighted assets with portfolio-level metrics

Example Strategies

Golden Cross:

"Backtest SMA 50/200 crossover on BTC/USDT daily for 2024"

RSI Mean Reversion:

"Buy when RSI(14) drops below 30, sell above 70 on EUR/USD daily with 5% stop loss"

MACD Crossover:

"Enter when MACD line crosses above signal line, exit on reverse, on ETH/USDT daily"

Portfolio:

"Backtest RSI strategy on a portfolio of 60% BTC and 40% ETH over 2024"

Optimization:

"Optimize the SMA crossover on BTC — try fast period 10-50 step 10, slow period 50-200 step 50"

Shareable Results

Every backtest gets a public URL you can share on Twitter, Discord, or with your trading community:

https://panther.watch/backtests/<backtest-id>

Results include performance metrics, equity curve chart, and full trade list.

Development

git clone https://github.com/PantherWatch/panther-mcp.git
cd panther-mcp
uv sync
PANTHER_API_KEY=pthr_... uv run panther-mcp

License

MIT