> ## Documentation Index
> Fetch the complete documentation index at: https://docs.nullboard.xyz/llms.txt
> Use this file to discover all available pages before exploring further.

# Usage Guide

> Learn how to capture strokes and configure training.

## Capturing Strokes

`Ink` provides a stroke management system that easily integrates with standard HTML5 Canvas events.

```typescript theme={null}
const canvas = document.getElementById('myCanvas');
const ink = new Ink();
let isDrawing = false;

canvas.addEventListener('mousedown', (e) => {
  isDrawing = true;
  // Start tracking a new stroke
  const stroke = ink.startStroke();
  stroke.addPoint(e.offsetX, e.offsetY);
});

canvas.addEventListener('mousemove', (e) => {
  if (!isDrawing) return;
  // Add points to the active stroke
  const strokes = ink.getStrokes();
  const currentStroke = strokes[strokes.length - 1];
  currentStroke.addPoint(e.offsetX, e.offsetY);
  
  // (Optional) Draw to canvas context here for visual feedback
});

canvas.addEventListener('mouseup', () => {
  isDrawing = false;
});
```

## Auto-Training Options

You can customize how the library trains itself via the `InkOptions` object passed to the constructor.

```typescript theme={null}
const ink = new Ink({
  auto: true,                // Enable auto-training (default: true)
  autoEpochs: 20,            // Increase epochs for potentially better accuracy
  autoAugmentFactor: 5,      // Generate more variations of sample data
  onTrainProgress: (epoch, total, logs) => {
    console.log(`Training progress: ${Math.round(epoch / total * 100)}%`);
  }
});
```

<Tip>
  **Performance Note**: Auto-training happens on the main thread in this version. For complex applications, consider running training during an initialization phase or loading screen.
</Tip>