Quick answers.

Yes. Both email + password and Google OAuth work. We recommend Google for the smoother re-login experience on mobile.

Voice transcripts, sessions, plans, and wellbeing data are permanently deleted within 30 days. We retain anonymised aggregate counts for product analytics, with no link back to you.

Technically yes, but the AI plans are personalised to one aspirant's stress, sleep, and progress patterns. Shared accounts produce worse plans for both of you. Free tier exists so each of you has your own.

Premium includes daily AI-generated plans, the Catch-Up Card, anomaly detection in your study patterns, the wellbeing dashboard, and ad-free experience. Free tier keeps voice logging, history, streaks, and the progress road — forever.

India-first pricing. ₹99/year is roughly $1.20 — well below international SaaS pricing. We use Razorpay for INR and Stripe for USD. If you can pay in INR, you should — better experience and lower price.

Cancel within 7 days of any annual payment for a full refund. After 7 days, prorated refunds via support. See /refund for full policy.

No. Parental mode is free. We don't think family safety should be a paid feature.

In your browser, via Whisper running on WebAssembly. The audio never reaches our servers. On older mobile devices that can't run WASM, we fall back to Groq's Whisper API — and in that case we delete the audio immediately after transcription.

Yes. Stress, mood, and sleep fields are encrypted at the application layer with AES-256-GCM before they hit the database. Even a database breach wouldn't expose them in cleartext.

No. Parental mode is read-only on progress percentages, session counts, and burnout status. It never exposes voice transcripts, journal entries, or stress level numbers.

Supabase managed PostgreSQL, region ap-south-1 (Mumbai). All data resides in India.

Groq's Llama 3.1 70B is the primary model. If Groq is slow or down, we fall back to a rule-based planning engine that runs locally on our server. Plans always generate — they never fail silently.

Maximum 3 LLM calls per user per day: one plan generation at 11pm, plus up to two on-demand insights. This caps your privacy exposure and keeps costs predictable.

Yes. Every plan has a 'why this plan' link that shows the inputs the AI received — yesterday's sessions, your stress signals, your progress percentages. No black box.

Generate a one-time link from your settings page, share it with them, they verify their email, and they're in. The link expires in 24 hours after creation.

Yes, in two clicks from /settings/parental. Revocation is instant — the dashboard goes dark immediately for them.

They'll need their own. We use email as the unique identifier for read-only access, and we don't allow the same email on both sides of the parental link.