Inspiration

The inspiration for SafeMeets came from the rising concerns around gender-based violence (GBV) and the risks involved when meeting strangers for the first time whether for professional, casual, or transactional purposes. We wanted to create an app that not only verifies the identity of the person you’re meeting but also provides an emergency panic button for instant help in unsafe situations

What it does

SafeMeets allows users to: Submit their details and upload verification images from different orientations (front, left, right). Use AI to confirm that images are clear and correctly oriented. Store verification data securely in the cloud. Trigger a panic button that instantly alerts a pre-set emergency contact if they feel unsafe.

How we built it

Frontend: Built using Lovable.dev for rapid prototyping and UI design. Backend: Hosted on AWS using API Gateway and Lambda for processing requests. Image Storage: Stored securely in AWS S3. AI Integration: Used AWS Rekognition to detect facial orientation and ensure correct image submission. Panic Button: Sends immediate notifications through the backend to the chosen emergency contact.

Workflow: User Uploads Image → Frontend (Lovable.dev) → API Gateway → Lambda Processing → S3 Storage User Presses Panic Button→Frontend Alert→API Gateway→Lambda Notification to Contact

Challenges we ran into

CORS errors due to dynamic frontend URLs from Lovable.dev. Base64 image payload issues causing Lambda parsing failures. Tuning AI accuracy to avoid false orientation matches. Ensuring real-time panic alerts with minimal delay. Staying within AWS Free Tier limits for all services.

Accomplishments that we're proud of

Successfully integrated AI-powered verification with orientation detection. Built a fully working panic button with instant emergency notifications. Designed a clean, functional UI using Lovable.dev without manual CSS heavy lifting. Ensured secure data handling between frontend and backend.

What we learned

How to integrate AI services like AWS Rekognition into real-world workflows. Best practices for CORS configuration in dynamic frontend environments. Efficient handling of image data in JSON payloads. The importance of real-time feedback for safety-critical features like the panic button.

What's next for SafeMeets

Adding location tracking to panic alerts for faster response. Integrating SMS/WhatsApp API for multi-channel emergency notifications. Expanding AI checks to include identity document verification.

Offering multi-language support for global accessibility.

Built With

  • amazon-web-services
  • and-ai-powered-image-analysis.-data-handling-includes-base64-encoding
  • and-rekognition-for-storage
  • apis
  • aws-api-gateway
  • aws-dynamodb
  • aws-lambda
  • aws-rekognition
  • base64-encoding
  • cors
  • javascript
  • lovable
  • lovable.dev
  • python
  • s3
  • typescript
Share this project:

Updates

Submission history