dynamic_signature_pad

Capture handwritten signatures as dynamic time-series data — vector strokes with velocity and optional pressure — not just a flat image. Lossless JSON and SVG export, plus PNG rendering. Pure Dart/Flutter, no native code and no third-party dependencies.

The data model is designed to map conceptually to ISO/IEC 19794-7 (signature/sign time series data), so a signature is preserved as the dynamics of how it was drawn — useful both for crisp reproduction and for behavioural (dynamic) signature analysis.

Features

Dynamic capture: each point keeps x, y, t and (when available) pressure; velocity is computed.
Velocity-based variable stroke width for realistic ink.
Optional full-screen, orientation-locked capture screen (showDynamicSignature) for maximum signing width without fighting the OS auto-rotate; it restores the orientation on exit.
Lossless outputs: JSON (canonical), SVG (vector). PNG for embedding.
Smooth & light: point decimation keeps capture cheap on mid-range devices.
Finger and stylus; clear / undo.
Tiny, stable public API; capture only — persistence is up to the host.

Install

dependencies:
  dynamic_signature_pad: ^0.0.1

Usage

// Full-screen, landscape-locked capture (recommended for max signing space).
// Locks landscape while signing and restores orientation on exit.
final result = await showDynamicSignature(context, title: 'Sign here');
if (result != null) {
  result.strokeJson; // lossless dynamic data (evidence)
  result.svg;        // vector
  result.pngBytes;   // raster (e.g. to embed in a PDF)
}

// Or embed the pad inline (give it a bounded box; ~2:1 is a good aspect):
final controller = DynamicSignatureController(
  config: const DynamicSignaturePadConfig(showGuideLine: true),
);

AspectRatio(aspectRatio: 2, child: DynamicSignaturePad(controller: controller));

// Read the inline controller when done:
await controller.toResult(); // null if empty
controller.toJson();
controller.toSvg();
await controller.toPng(pixelRatio: 3.0);
controller.clear();
controller.undo();

Output formats

Format	Method	Use
JSON (stroke data)	`toJson()`	Canonical, lossless — the dynamic evidence
SVG	`toSvg()`	Vector render, scales to any size
PNG	`toPng()`	Raster, DPI-aware (embedding/compatibility)
All three	`toResult()`	The standardized `SignatureResult`

Performance

You capture a point model in memory; SVG/PNG are derived once at the end. Capturing dynamic points costs about the same as capturing a plain point list; the only extra live cost is drawing the realistic (variable-width) stroke, which is bounded by signature length and kept cheap via point decimation and a RepaintBoundary.

Roadmap

ISO/IEC 19794-7 binary exporter.
Quadratic-bezier smoothing; cached layer for committed strokes.
Web/desktop validation; optional dynamic signature matching.

License

MIT — see LICENSE.