AI-Driven Task Manager MVP

I’m ready to move from architecture diagrams to a live, web-based MVP that automates and tracks user-submitted tasks. The stack is already chosen—Next.js on the front end, Supabase for auth, database, and storage, Vercel for hosting. Your mission is to stitch these pieces together and deliver a production-ready app.

Core build
• Responsive dashboard in Next.js that shows task status and history.
• Supabase set up for basic email/password login (no social auth or MFA yet), PostgreSQL tables, and file storage.
• Gmail API integration so the system can watch an inbox, parse new emails, and hand off each message to the pipeline.
• OpenAI API call that classifies the parsed task, then simple routing logic to decide if it’s automated or requires a human touch.
• Stripe checkout to let me turn on paid tiers from day one.
• Deployment to Vercel with environment variables locked down and data handled securely end to end.

Deliverable checklist
– Authenticated web app available on my Vercel account
– Email ingestion running and visible in Supabase logs
– Task objects written to the database with an AI-generated “category” field
– Dashboard pages showing live status updates and manual actions when needed
– Stripe billing flow tested in both sandbox and live mode

You’ll collaborate directly with me; I have wireframes, a schema draft, and API keys ready. Clean code, clear commits, and brief Loom walkthroughs of major features will speed up acceptance.