QuoteForge

QuoteForge is a contract-manufacturing quote and order tracking demo built with Django REST Framework, React + Redux Toolkit, PostgreSQL, Docker Compose, GitHub Actions, and Terraform for AWS foundations.

Demo

▶️ Watch the QuoteForge demo

Run locally

docker-compose up --build

The client runs at http://localhost:5173, API at http://localhost:8001/api, and Django admin at http://localhost:8001/admin. Create a user with docker-compose exec backend python manage.py createsuperuser; set is_staff to use the staff-only order advance endpoint. Obtain a JWT at POST /api/auth/token/ and place its access value in browser localStorage as access for the demo UI.

API flow

• POST /api/estimate/: creates a Part and Quote, returning price/lead time.
• POST /api/quotes/:id/accept/: accepts a quote and opens an Order.
• POST /api/orders/:id/advance/: staff-only, follows quoted → in_production → quality_check → shipped → delivered and creates an audit event.
• GET /api/quotes/, GET /api/orders/: customer-scoped unless the authenticated user is staff.

Development and verification

cd backend && USE_SQLITE=true python manage.py test
cd frontend && npm install && npm run build
python scripts/generate_synthetic_data.py
python scripts/train_estimator.py

The GitHub Actions workflow runs backend tests and a frontend production build. Terraform in infra/terraform provisions an AWS VPC, two subnets, RDS Postgres, and an ECS cluster; run terraform init, review terraform plan, then apply with a real secret via TF_VAR_db_password.

Architecture

React keeps draft quote form state in Redux. Django owns data, authorization, estimation and status transitions. PostgreSQL stores customers, parts, quotes, orders, suppliers and immutable status events. A rule estimator is the safe default; an optional trained linear model is loaded only after the supplied training script produces it. See MODEL.md for scope and limitations.