cloudQuant/backtrader-mcp

Overview

Local-first MCP server that lets AI assistants develop Backtrader strategies. Provides 30 typed tools for managing immutable datasets, private drafts, and running bounded subprocess backtests. Works offline and is backtest-only.

Best for

Best for
AI-assisted Backtrader strategy developers who want offline, sandboxed backtesting

Use cases

• Develop and iterate on Backtrader strategies with an AI assistant
• Run backtests in a sandboxed subprocess without network access
• Manage versioned datasets and private strategy drafts locally

How to use

Tested with

Claude Desktop, Claude Code, ChatGPT

Notes

Local-first MCP server that lets AI assistants develop Backtrader strategies. Provides 30 typed tools for managing immutable datasets, private drafts, and running bounded subprocess backtests. Works offline and is backtest-only.

1 stars on GitHub. Last updated 2026-09-20. Licensed MIT.

Use cases

• Develop and iterate on Backtrader strategies with an AI assistant
• Run backtests in a sandboxed subprocess without network access
• Manage versioned datasets and private strategy drafts locally

Pros

• Local-first design keeps data private and offline
• Bounded subprocess backtests prevent runaway execution
• Typed tools give AI assistants a clear interface

Cons

• Only 1 star, early stage
• Backtest-only, no live trading support
• Requires Python environment and MCP client setup

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.