CookSnap

CookSnap is a small full-stack project that detects ingredients from images and helps users manage recipes, pantry items, and favorites.

Quick Overview

• Backend: Node.js + Express API (in backend/).
• Detector: Python image-detector service (in detector/).
• Frontend: Vite + React app (in frontend/).

Run (development) This repository uses pnpm workspaces. From the project root you can install dependencies and run services.

Install all workspace dependencies:

pnpm install

Run all services in parallel (development):

pnpm dev

Run services individually from the project root:

• Backend: pnpm dev:backend (or pnpm -F backend dev)
• Detector: pnpm dev:detector (use pnpm -F detector install first to create the Python venv)
• Frontend: pnpm dev:frontend (or pnpm -F frontend dev)

Files & structure

• backend/ — API server, routes, models, and controllers.
• detector/ — image detector service (PyTorch model + FastAPI wrapper).
• frontend/ — React UI (Vite).

Next steps

• See backend/README.md, detector/README.md, and frontend/README.md for per-service instructions and env variables.

---

Development Notes (pnpm workspace)

Service URLs (development defaults):

• Frontend: http://localhost:5173 (Vite dev server)
• Backend API: http://localhost:3000
• Detector: http://localhost:8000 (POST /detect)

Common commands (run from repository root):

pnpm install            # install workspace deps
pnpm dev                # run all dev services in parallel
pnpm build              # build frontend
pnpm start              # run start scripts in parallel (preview/server)

pnpm dev:frontend       # run frontend only
pnpm dev:backend        # run backend only
pnpm dev:detector       # run detector only

Notes:

• The detector uses Python and creates a virtual environment when you run pnpm -F detector install.
• Environment variables for each service are documented in the service READMEs.
• If you still want to run services individually without pnpm you can follow the per-service README instructions.
• You need a YOLO model.pt file for the model to detect images otherwise it runs on a mock mode.