Marg

AI-guided brand kits and proposals. Describe a deal in plain language, and Marg drafts a full, on-brand proposal — packages, timeline, terms — ready to send as a single shareable link.

Stack

• Next.js 16 (App Router, Turbopack) + React 19 + TypeScript
• Supabase — Postgres, Auth, Row Level Security, Storage
• AI SDK + OpenAI (gpt-4o) for proposal generation and the intake chat
• Firecrawl — scrapes a client's site to read their brand (colors, fonts, logo)
• Resend — transactional email; Upstash Redis — AI-endpoint rate limiting
• Stripe — billing (paused — see docs/DECISION_LOG.md)
• Sentry — error tracking
• Deployed on Vercel

Getting started

1. Install dependencies

   npm install
   

2. Configure environment variables — copy .env.example to .env.local and fill in the required values (Supabase, OpenAI, Firecrawl). Everything else in that file is optional for local dev and degrades gracefully when unset (see the comments in .env.example).
3. Start a local Supabase stack (requires the Supabase CLI and Docker):

   npx supabase start
   

   This applies every migration in supabase/migrations/ and seeds local data from supabase/seed.sql. Point .env.local's Supabase variables at the local URL/keys it prints.
4. Run the dev server

   npm run dev
   

   Open http://localhost:3000.

Scripts

Command	What it does
npm run dev	Start the Next.js dev server (Turbopack)
npm run build	Production build — also runs Next.js's own TypeScript check
npm start	Serve a production build
npm run lint	ESLint
npm test	Unit/integration tests (Vitest)
npm run test:e2e	End-to-end tests (Playwright)

CI (.github/workflows/ci.yml) runs typecheck, build, test, and lint on every pull request.

Project docs

docs/ has the deeper reference material: PRD.md (product spec), SCHEMA.md (database schema), ROADMAP.md, DECISION_LOG.md (why things are the way they are, including paused or reversed decisions), and TEST_PLAN.md.