Optara (Desktop + API + AI)

Optara uses three layers:

1. Backend (Django API only)
   • Stores users, encrypted embeddings, recognition logs
   • Runs at http://127.0.0.1:8000/api/
2. Desktop App (CustomTkinter)
   • Live camera feed
   • Face mapping workflow (capture -> prompt for name -> register)
3. AI Engine (service layer)
   • Face detection + embedding extraction + FAISS search

Face Mapping Flow (requested behavior)

1. Start camera in desktop app.
2. Click Map Face.
3. App captures multiple face crops (12 samples by default).
4. After capture completes, app prompts for a name.
5. Name + face images are sent to POST /api/register/ and saved.

Encryption

Embeddings are encrypted at rest in the database using Fernet (from cryptography).

• Set a strong key in your environment:

export OPTARA_ENCRYPTION_KEY="replace-with-a-long-random-secret"

If not set, a development default is used (not safe for production).

Common Error: WinError 10061

Connection refused usually means backend is not running.

Start backend first:

cd backend
python manage.py runserver

Then run desktop app:

python desktop_app/main.py

API Endpoints

• POST /api/register/
• POST /api/recognize/
• GET /api/logs/

Install

python -m venv .venv
.venv\Scripts\activate
pip install -r requirements.txt