Stripe Subscriptions & Voice Mode Upgrade

Our Mentora.ai MVP already handles chat, voice input/output, quizzes, and “upload & learn.” What’s missing is a production-ready billing flow and a true hands-free conversation mode. I want a developer who can extend the existing Next.js / React / Node stack (MongoDB + Supabase for auth) without breaking a thing.

Feature 1 – Stripe Subscription (Web only)
Stripe Checkout must offer three plans—Basic, Standard, and Premium—each available monthly or yearly. The pricing page needs to show straightforward plan features and a clear cost breakdown; testimonials can wait for marketing. After payment, webhooks have to update the user’s record so premium features unlock instantly and downgrade or cancellation is reflected just as fast. Success and cancel screens should feel native to the current UI.

Feature 2 – Hands-Free AI Voice Conversation
The core pipeline (recording → STT → LLM → TTS) already works. What I still need is continuous mic streaming via WebSocket, live captions that appear mid-sentence and finalize on pause, real-time TTS playback, and classic barge-in so the bot stops talking the moment the user starts. All of this has to run smoothly inside the existing React component tree and respect our current latency budget.

Deliverables
• Fully wired Stripe Checkout with working webhooks, plan enforcement, and DB updates
• Responsive pricing, success, and cancel pages matching current design system
• WebSocket-based audio stream with partial/final captions and barge-in logic
• Integrated, low-latency TTS playback that coexists with our LLM calls
• Clean, documented PR plus a short Loom walkthrough

Acceptance
I’ll run end-to-end tests: subscribe, renew, cancel, and confirm database flags update; then conduct a 5-minute continuous voice chat where captions, TTS, and barge-in behave exactly as specified. If both flows pass without regressions, the milestone is complete.