Tech Stack
Technologies, frameworks, and infrastructure used in CROP
Tech Stack
| Layer | Technology | Used In |
|---|---|---|
| Frontend | Next.js 16, React 19, TailwindCSS v4 | CROP Front, Admin, Form |
| Auth | Clerk | All apps and services |
| Backend | Bun, Hono | Parts Services, Delivery Services |
| AI/ML | FastAPI (Python), PyTorch | PDF Parser, Schematics, CLIP |
| Databases | MongoDB Atlas, Typesense, Weaviate | Search, catalog, vectors |
| Payments | Stripe | Payment Service |
| Shipping | UPS API | Delivery Services |
| Infrastructure | GCP Cloud Run, Cloudflare, Vercel | All services |
| CI/CD | GitHub Actions | All repositories |
| Monitoring | Sentry, GCP Logging | Frontend + Backend |
| Design | Figma, shadcn/ui | CROP Front |
Key Decisions
- Bun over Node.js: Faster startup, native TypeScript, built-in test runner
- Hono over Express: Lightweight, edge-compatible, type-safe routing
- Typesense over Elasticsearch: Simpler ops, typo tolerance, instant search
- Weaviate: Vector database for RAG and semantic search
- Cloudflare Workers: Edge deployment for CROP Front (production)
- Vercel: Dev/staging deployments for frontend apps
- MongoDB Atlas: Primary data store, flexible schema for parts data