📸 Capture & Match
- ID + selfie capture via device camera.
- Face match between ID face and live selfie.
- Legal-age decision (18/21/custom) with clear messaging.
ID & Face Verification
Legal Age — ID & Face Verification (Flutter)
Legal Age is a modern, agency-ready mobile app to verify legal age and confirm identity by comparing an ID photo to a live selfie. The app supports dual cloud providers (AWS + Azure) with automatic failover, keeping secrets on the server side for security and compliance.
Features
Everything you need for age & ID verification in one Flutter app.
🤖 Cloud Intelligence
- AWS Rekognition:
CompareFaces&DetectFaces. - Azure Face API:
verify&detect?returnFaceAttributes=age. - Age estimation with confidence scores.
🧠 Automatic Failover
- Primary provider preference (AWS or Azure).
- Automatic fallback on network/HTTP/parse errors.
- UI shows which provider handled the final result.
🎨 Agency-Ready UI
- Material 3, high contrast, dark mode.
- Step-by-step flows with clear statuses.
- Results summaries and explanatory text for edge cases.
⚙️ Runtime Settings
- Switch primary provider at runtime.
- Override API bases/keys and auth headers.
- Configure legal-age threshold; persisted via SharedPreferences.
🔐 Security First
- Server-side secrets; app only talks to your proxies.
- Optional auditing hooks (e.g., Firestore logs).
- Clear guidance for JWT protection and data retention.
Architecture Overview
Clean separation between Flutter app, proxy services, and cloud providers.
Mobile (Flutter)
main.dart— app entry, global error handling, theming.id_scanner_app.dart— shell with a private navigator.id_scanner_home_page.dart— home with quick actions.main_screen.dart— guided overview and entry to verification.face_verification_screen.dart— capture, verify, results.settings_screen.dart— provider, threshold, and runtime overrides.
Backends (Server-Side)
- AWS Lambda + API Gateway for Rekognition.
- Azure Functions for Azure Face API.
- Both expose normalized JSON for the Flutter app.
Data Flow
Camera (ID & Selfie) → Flutter App → Your Proxy API (AWS or Azure). On failure,
the orchestrator falls back to the secondary provider and returns normalized results:
{ matched, confidence, estimatedAge, ageConfidence, provider }.
Getting Started
Prerequisites
- Flutter 3.19+.
- Xcode (iOS) and Android SDK (Android).
- AWS and/or Azure accounts for backends.
Core Dependencies
dependencies:
flutter:
sdk: flutter
http: ^1.2.2
image_picker: ^1.1.2
google_fonts: ^6.2.1
shared_preferences: ^2.3.2
Runtime Configuration
Environment-agnostic builds using --dart-define and runtime overrides.
| Dart Define | Example | Purpose |
|---|---|---|
AWS_API_BASE |
https://abc.execute-api.us-east-1.amazonaws.com/prod |
API Gateway base URL for AWS Lambdas. |
AZURE_FN_BASE |
https://your-fn.azurewebsites.net |
Azure Function App base URL. |
AZURE_FN_CODE |
... (optional) |
Function key if required by your auth level. |
AUTH_BEARER |
eyJhbGciOiJI... |
JWT/ID token header for your gateways (optional). |
PRIMARY_PROVIDER |
aws or azure |
Default provider preference. |
All of these can be overridden at runtime via the Settings screen and persisted using
SharedPreferences.
Platform Permissions
iOS — Info.plist
<key>NSCameraUsageDescription</key>
<string>Camera is used to capture ID photos and selfies for verification.</string>
<key>NSPhotoLibraryUsageDescription</key>
<string>Photo library access is used to pick images for verification.</string>
Deploying the Backends
AWS + Azure patterns ready for your infrastructure team.
AWS: Rekognition + Lambda + API Gateway
IAM Role policy (Lambda):
{
"Version": "2012-10-17",
"Statement": [
{"Effect":"Allow","Action":["rekognition:CompareFaces","rekognition:DetectFaces"],"Resource":"*"},
{"Effect":"Allow","Action":["s3:GetObject"],"Resource":"*"}
]
}
- Create two Node.js 20 Lambdas:
POST /compare,POST /detect. - Wire them via an HTTP API in API Gateway with CORS enabled.
- Use multipart form-data:
idImage,selfieImage, orimage.
Sample tests:
curl -X POST "$AWS_API_BASE/compare" \
-F idImage=@id.jpg -F selfieImage=@selfie.jpg
curl -X POST "$AWS_API_BASE/detect" \
-F image=@selfie.jpg
Build & Run
Development
flutter pub get
flutter run \
--dart-define=AWS_API_BASE=https://abc.execute-api.us-east-1.amazonaws.com/prod \
--dart-define=AZURE_FN_BASE=https://your-fn.azurewebsites.net \
--dart-define=PRIMARY_PROVIDER=aws
Android Release
flutter build apk --release \
--dart-define=AWS_API_BASE=... \
--dart-define=AZURE_FN_BASE=... \
--dart-define=PRIMARY_PROVIDER=aws
# or:
# flutter build appbundle --release ...
iOS Release
flutter build ios --release \
--dart-define=AWS_API_BASE=... \
--dart-define=AZURE_FN_BASE=... \
--dart-define=PRIMARY_PROVIDER=azure
# Then archive/sign in Xcode as usual
Testing & QA
Happy Path
- Clear ID and selfie of the same person →
matched=true, high confidence. - Plausible age from selfie with confidence metrics.
Fallback Scenarios
- Temporarily disable one backend → expect successful result via the other.
- UI clearly indicates “Fallback used”.
Edge Cases
- No face detected →
matched=false,estimatedAge=null. - Multiple faces → first face used; prompt user to retake.
- Poor lighting / occlusions → lower confidence but safe messaging.
Troubleshooting
- AWS_API_BASE / AZURE_FN_BASE empty: Set via
--dart-defineor Settings. - 403/401: Check
AZURE_FN_CODEor JWT auth. - CORS errors: Enable CORS on API Gateway / Function App (for web).
- Camera denied: Verify permissions and in-app prompts.
- No provider available: Both failed; check logs and network.
- Use Settings “Ping” buttons to test reachability quickly.
Interested in acquiring Legal Age?
Share a bit about your organization, compliance requirements, and how you plan to use Legal Age. We’ll follow up with a detailed technical brief and next steps.