unified_face_camera #

Production-ready face capture for Flutter — a drop-in camera widget with real-time face alignment, liveness anti-spoofing, smart quality gates, and tamper-evident timestamp embedding. Built for KYC, attendance, identity verification, and any workflow that needs trustworthy face photos.

One widget. Offline ML Kit detection. Native timestamp overlay. No server required for liveness checks.

Why unified_face_camera? #

Most camera plugins give you a preview. This one gives you a validated capture pipeline:

Capability	What you get
Face alignment	Live overlay + guided feedback ("move closer", "look straight", "blink")
Quality gates	Distance, head pose, eye contour, and blink checks before shutter unlocks
Liveness	Passive anti-spoofing score on every frame — blocks photo-of-a-photo attacks
Smart capture	Shutter stays disabled until all checks pass on a fresh frame
Timestamp proof	Native overlay on the saved image (date, time, optional GPS)
Cross-platform	Android & iOS with a single Dart API

Features #

Real-time ML Kit face detection with animated overlay
Passive liveness via face_anti_spoofing_detector
Blink detection to confirm a live subject
Head pose validation (yaw, pitch, roll limits)
Distance checks — face must fill the frame correctly
Front / back camera switch with cooldown re-validation
Flash modes (off, auto, always, torch)
Portrait-locked capture with EXIF rotation fix
Optional GPS geotag burned into the image via native code
Static permission helpers — checkPermission() / requestPermission()

Quick start #

1. Add the dependency #

dependencies:
  unified_face_camera: ^0.0.1

2. Request permission & show the widget #

import 'package:flutter/material.dart';
import 'package:unified_face_camera/unified_face_camera.dart';

class FaceCapturePage extends StatelessWidget {
  const FaceCapturePage({super.key});

  @override
  Widget build(BuildContext context) {
    return Scaffold(
      body: UnifiedFaceCamera(
        useFrontCamera: true,
        onCapture: (path) {
          // Final image path — timestamp (and optional GPS) already applied
          debugPrint('Captured: $path');
          Navigator.of(context).pop(path);
        },
        onError: (error) => debugPrint('Error: $error'),
        onClose: () => Navigator.of(context).pop(),
      ),
    );
  }
}

3. Check permissions first (recommended) #

final granted = await UnifiedFaceCamera.requestPermission();
if (!granted) {
  // Show a dialog explaining why camera access is needed
  return;
}

How it works #

┌─────────────┐     ┌──────────────────┐     ┌─────────────────┐
│ Camera feed │ ──► │ ML Kit detection │ ──► │ Quality + blink │
└─────────────┘     └──────────────────┘     │     checks      │
                                              └────────┬────────┘
                                                       │
                       ┌──────────────────┐            ▼
                       │ Anti-spoof score │ ◄── Liveness gate
                       └────────┬─────────┘
                                │ pass
                                ▼
                       ┌──────────────────┐     ┌─────────────────┐
                       │  User taps       │ ──► │ Native timestamp│
                       │  shutter         │     │ + optional GPS  │
                       └──────────────────┘     └─────────────────┘

Camera streams frames to ML Kit for face detection.
The view model validates distance, pose, contours, and blink.
Anti-spoofing runs on the face crop — capture unlocks only above threshold.
On shutter press, the image is rotated, timestamped natively, and returned via onCapture.

Platform setup #

Android #

Minimum SDK 24 in android/app/build.gradle.kts:

defaultConfig {
    minSdk = 24
}

ML Kit face model — add inside <application> in AndroidManifest.xml:

<meta-data
    android:name="com.google.mlkit.vision.DEPENDENCIES"
    android:value="face" />

Camera and location permissions are merged by the plugin. Request location at runtime if you want GPS on the timestamp.

iOS #

Add to Info.plist:

<key>NSCameraUsageDescription</key>
<string>We need camera access to capture your face photo for verification.</string>

<!-- Optional — only if you want GPS on the timestamp -->
<key>NSLocationWhenInUseUsageDescription</key>
<string>We use your location to embed GPS coordinates in captured photos.</string>

Minimum iOS version: 13.0.

API reference #