NAAb Pivot - Recording Demos for GitHub

March 1, 2026 · View on GitHub

This guide shows how to create visual demos for your README.

Quick Demo

cd demos
./pivot-demo.sh

This runs an automated demo showing Pivot optimizing Python code to Go (3.5x speedup).

Recording Options

Option 1: Screenshot (Easiest) ⭐

Best for: Quick visual proof, no tools needed

  1. Run the demo:

    ./pivot-demo.sh

  2. When you see the performance comparison, screenshot it:

    • Termux: Power + Volume Down
    • Desktop: Your OS screenshot tool

  3. Add to README:

    ![Pivot Performance](demos/performance.png)

Recommended screenshots:

  • The performance comparison bar chart
  • The "3.5x faster" speedup summary

Option 2: Terminal Recording with asciinema

Best for: Embeddable terminal player

  1. Install asciinema:

    pkg install asciinema

  2. Record:

    asciinema rec pivot-demo.cast
./pivot-demo.sh
# Press Ctrl+D when done

  3. Upload:

    asciinema upload pivot-demo.cast

  4. Add to README:

    [![asciicast](https://asciinema.org/a/YOUR_ID.svg)](https://asciinema.org/a/YOUR_ID)

What to Capture

Key Moments:

  1. Analysis Phase

    • Complexity scoring
    • Optimization recommendations

  2. Performance Comparison (most important!)

    • Python: 2,843 ms
    • Go: 812 ms
    • Visual bar chart
    • "3.5x faster" highlight

  3. Parity Validation

    • Green ✓ marks
    • "99.99% confidence"
    • 100 test cases passing

Adding to README

Example 1: Before/After

## Performance Improvement

**Before (Python):**
```python
def heavy_computation(n):
    result = 0.0
    for i in range(n):
        result += (i ** 2) ** 0.5
    return result
# Time: 2,843 ms

After (Go):

func heavyComputation(n int) float64 {
    result := 0.0
    for i := 0; i < n; i++ {
        result += math.Sqrt(math.Pow(float64(i), 2))
    }
    return result
}
// Time: 812 ms (3.5x faster!)

Performance Comparison


### Example 2: Visual Results
```markdown
## Demo Results

| Analysis | Optimization | Performance |
|----------|--------------|-------------|
| ![Analysis](demos/analysis.png) | ![Code](demos/optimized.png) | ![Benchmark](demos/benchmark.png) |

**Result: 3.5x speedup with proven correctness (99.99% confidence)**

Tips for Best Results

  • Highlight the speedup: The performance comparison is the money shot
  • Show the bar chart: Visual comparison is very effective
  • Include parity validation: Proves correctness
  • Terminal width: ~80 columns for best display

Editing the Demo

The demo script is at demos/pivot-demo.sh. You can:

  • Change speedup numbers to match your real results
  • Adjust target language (Go, Rust, C++)
  • Add more benchmark comparisons
  • Customize timing

Using Real Benchmark Data

To show actual performance improvements:

  1. Run Pivot on real code:

    ./naab/build/naab-lang pivot.naab evolve slow.py

  2. Capture the benchmark output

  3. Update demo script with real numbers

  4. Screenshot the actual results