mcpbr
Benchmark Your MCP Server
Stop guessing if your MCP server actually helps. Get hard numbers comparing tool-assisted vs. baseline agent performance on real tasks.
$ pip install mcpbr && mcpbr init && mcpbr run -c mcpbr.yaml -n 1 -v
What You Get
Evaluation Results
Summary
+-----------------+-----------+----------+
| Metric | MCP Agent | Baseline |
+-----------------+-----------+----------+
| Resolved | 8/25 | 5/25 |
| Resolution Rate | 32.0% | 20.0% |
+-----------------+-----------+----------+
Improvement: +60.0%
Benchmark Categories
- Code Generation 2 benchmarks
- Code Understanding 1 benchmark
- Knowledge & QA 4 benchmarks
- Math & Reasoning 3 benchmarks
- ML Research 1 benchmark
- Security 1 benchmark
- Software Engineering 7 benchmarks
- Tool Use & Agents 6 benchmarks
Documentation
-
Installation
Prerequisites, install methods, and setup
-
Configuration
YAML config reference and examples
-
CLI Reference
All commands and options
-
About
Origin story, philosophy, and vision
Benchmark Your MCP Server
Get hard numbers comparing tool-assisted vs. baseline agent performance on real tasks.
Get Started Browse BenchmarksCreated by Grey Newell