tools-only Tutorials
Learn mcpbr through interactive, hands-on tutorials that guide you step by step.
Available Tutorials
| Tutorial | Difficulty | Time | Description |
|---|---|---|---|
| Getting Started | Beginner | ~10 min | Install mcpbr, create your first config, and run a benchmark |
| Configuration | Intermediate | ~15 min | Master YAML configuration, environment variables, and validation |
| Benchmarks | Intermediate | ~15 min | Explore benchmark selection, filtering, and custom benchmarks |
| Analytics | Advanced | ~20 min | Use the analytics engine for trends, comparisons, and reports |
Using the Tutorial CLI
List tutorials
mcpbr tutorial list
Start a tutorial
mcpbr tutorial start getting-started
Check your progress
mcpbr tutorial progress
Reset and start over
mcpbr tutorial reset getting-started
Tutorial Workflow
Each tutorial consists of multiple steps. For each step you'll:
- Read the instruction and explanation
- Do the action (run a command, create a file, etc.)
- Validate — the tutorial checks your work automatically
- Continue to the next step
!!! tip "Hints"
Stuck on a step? Type hint when prompted to get a helpful nudge.
!!! info "Progress is saved"
Your progress is saved automatically in ~/.mcpbr_state/tutorials/. You can quit and resume any tutorial later with mcpbr tutorial start <id>.
What You'll Learn
Getting Started (Beginner)
- Installing mcpbr with pip or npm
- Creating a minimal configuration file
- Running your first benchmark evaluation
- Understanding the results output
Configuration (Intermediate)
- YAML configuration structure and fields
- Environment variable substitution (
${VAR}) - Configuration validation with
mcpbr config validate - Advanced options: timeouts, concurrency, resource limits
Benchmarks (Intermediate)
- Choosing the right benchmark for your use case
- Filtering tasks by difficulty, category, and tags
- Running multiple benchmarks in sequence
- Interpreting benchmark-specific metrics
Analytics (Advanced)
- Storing results in the analytics database
- Generating trend reports and detecting regressions
- Comparing models and MCP servers side-by-side
- Creating HTML, Markdown, and PDF reports