👋 Welcome to the mcpbr Community!

[greynewell]

Welcome to the mcpbr community! 👋

Whether you're building an MCP server, evaluating tool-use capabilities, or just exploring the ecosystem, we're glad you're here!

What is mcpbr?

mcpbr (Model Context Protocol Benchmark Runner) helps you measure whether your MCP server actually improves agent performance. No vibes, just data.

• 🎯 Controlled experiments: Same model, same tasks, only variable is your MCP server
• 📊 Real benchmarks: SWE-bench, CyberGym, TerminalBench (MCPToolBench++ coming soon!)
• 🐳 Reproducible: Docker containers with pinned dependencies

Quick Links

• 📚 Getting Started - Run your first benchmark in 5 minutes
• 🛠️ Contributing Guide - Help build the project
• 🗺️ Roadmap - See what's coming (200+ features planned!)
• ❓ FAQ - Common questions answered

Introduce Yourself!

Tell us about yourself:

• What brought you to mcpbr?
• What MCP server are you working on (or interested in)?
• What benchmarks are you most interested in?
• What features would you like to see?

How to Contribute

We have 30+ good first issues perfect for newcomers:

• https://github.com/greynewell/mcpbr/labels/good%20first%20issue

Or just:

• Share your benchmark results
• Report bugs
• Suggest features
• Improve documentation
• Star the repo ⭐

Looking forward to building together! 🚀