Homebrew

Homebrew is a web app for home cafés: small coffee setups run from someone's kitchen or garage. Customers find a café and pre-order drinks for pickup, café owners run their café, and Admins keep the catalogue and reviews trustworthy. Each role has one advisory AI feature, and every AI suggestion leaves the decision with a person.

• Customers browse and filter cafés; read menus, Singapore opening hours and public reviews; pre-order for pickup and pay at the counter; view and cancel orders; and review collected orders. Brew Finder suggests cafés and drinks from a natural-language search.
• Café owners apply to register a café; manage its profile, opening hours, pickup cutoff and menu; run the order queue and print today's prep sheet; and view sales. Prep Advisor suggests what to prepare more or less of on each weekday from the café's own sales.
• Admins approve or reject café applications, suspend or un-suspend cafés, moderate reviews and inspect the audit log. The Moderation Assistant suggests a label for each pending review; the Admin confirms every action.

Try it

Open the production app at https://homebrew-cs3227-prod.vercel.app and choose Try Homebrew → Create demo kit on Sign in. A kit creates a separate workspace with Customer, Café owner and Admin accounts, so you can test every role without help from the team. Save the credentials when they are shown: they appear only once. See the User Guide for details.

Getting started

Requires Node.js 24, Bun 1.4.3 and a running Docker engine for the local Supabase stack. From the repository root:

bun install --frozen-lockfile
bun run db:start
bun run db:reset
bun run dev

Then open http://127.0.0.1:3000. The app needs local configuration values before it starts; see the Developer Guide's setup section. Run bun run verify for lint, type checks, unit and integration tests, the spec and traceability checks, and a production build.

Documentation

• User Guide: accounts, demo kits and each role's features.
• Developer Guide: architecture, the spec-driven multi-agent workflow, security, testing and deployment.
• Product website
• Specifications and plan: workflow/specs/ and workflow/PLAN.md.
• Reflections: docs/Reflections.md.

Team: TaiaYovelaPang (Customer features), naa-siuuuu-ff (Café owner features) and shawnnygoh (Admin features).

Acknowledgements

Built with Next.js, React, Supabase, Zod, Tailwind CSS, Vitest, Playwright and Bun, hosted on Vercel. AI features use the SoC LLM service. See the Developer Guide's Acknowledgements for reuse from MP2 and other sources.