How swapnaniltripathy/memoryos works
MEMORYOS is a camera-based physical-memory prototype. A user enables the browser camera, captures a frame, and sends it to the frontend’s memory-frame API; the backend then detects objects, matches identities, and records their positions and movement in JSON-backed memory. Returned detections appear in the frontend with labels, IDs, and location. The graph follows the sampled runtime code; broader product capabilities are not documented in the provided README, and the database files are created at runtime rather than present in the tree.
Main entry points: server.py Memory-frame API · Camera.jsx Camera capture · App.jsx App shell · identity.py Object identity
Components 11
Camera interface
- React entry
frontend/src/main.jsx - App shell
frontend/src/App.jsx - Camera capture
frontend/src/Camera.jsx
Frame processing
- Memory-frame API
backend/server.py
Vision and memory
- Object detection
vision/detector.py - Object identity
vision/identity.py - Spatial memory
vision/spatial.py
Persistent memory
- Identity records
database - Spatial observations
database
Other components
- User
- Browser camera
Connections 11
- User → Camera capture: captures frame
frontend/src/Camera.jsx - Camera capture → Browser camera: requests stream
frontend/src/Camera.jsx - React entry → App shell: renders
frontend/src/main.jsx - App shell → Camera capture: renders
frontend/src/App.jsx - Camera capture → Memory-frame API: posts frame
frontend/src/Camera.jsx - Memory-frame API → Object detection: detects objects
backend/server.py - Memory-frame API → Object identity: matches identities
backend/server.py - Memory-frame API → Spatial memory: records positions
backend/server.py - Object identity → Identity records: reads and writes
vision/identity.py - Spatial memory → Spatial observations: reads and writes
vision/spatial.py - Memory-frame API → Camera capture: returns detections
frontend/src/Camera.jsx