💡 What Features Do You Want to See?

[greynewell]

What would you like to see in mcpbr? Help shape the future of MCP server benchmarking!

🗳️ Top Requested Features

Here are some ideas from our roadmap and community feedback. Which ones are most valuable to you?

Analytics & Insights

• Enhanced tool usage analytics (which tools get used, how often, success rates)
• Tool coverage reporting (identify unused tools)
• Error pattern analysis (categorize failures to improve MCP servers)
• Performance profiling (latency, token efficiency, overhead)

Benchmarks

• MCPToolBench++ - Comprehensive MCP tool use evaluation
• HumanEval - Code generation tasks
• MBPP - Python programming problems
• GAIA - General AI capabilities
• Custom benchmark support via YAML

Developer Experience

• Real-time dashboard during runs
• Interactive configuration wizard
• Cross-benchmark comparison reports
• Multi-benchmark runs in one command
• Results caching and resumption

Platform Support

• Multi-provider support (OpenAI, Gemini, Qwen)
• GitHub Action for CI/CD integration
• NPM package
• Homebrew formula
• Official Docker image

Output Formats

• CSV export
• YAML output
• XML export
• JUnit XML for CI
• Cost analysis in reports

💡 Your Ideas

What features are not on this list that you'd love to see?

🎯 What Matters Most?

What would make mcpbr most valuable for your work? Is it:

• More benchmarks?
• Better analytics?
• Easier setup?
• Faster execution?
• Better reports?
• Something else?

Let us know in the comments! Your feedback directly influences our roadmap priorities.