Privacy

Darshan (krishnadarshan.live) is a voice-call experience where you can speak with an AI persona of Krishna grounded in the Bhagavad Gita. This page explains what we collect, how long we keep it, and who we share it with. We've tried to keep it short and honest.

• Account info from Google.When you sign in with Google we receive your email, name, and profile picture. That's the minimum Google sends; we don't request anything else.
• Voice during a call. Your microphone audio is sent to our speech-to-text provider (Groq) for the duration of each utterance. The audio is not stored on our servers.
• Conversation transcripts (in memory). Krishna needs short-term memory to follow the thread of a conversation. Transcripts live in server memory only and are discarded when the call ends.
• Call metadata.Start/end timestamp, duration, turn count, and detected language. Used for billing minutes and showing your past darshans on the "Past darshans" page.
• Payment records. If you submit a UPI payment, we store the plan you chose, the amount, and the UPI transaction ID you entered. We do not see or store your bank or UPI account details — only the public transaction ID.
• Analytics. Standard Google Analytics 4 events (page views, button taps, call lifecycle). Tied to your Supabase user ID so we can analyze funnels but never to your name or email.

• Audio recordings — your voice is never persisted.
• Camera or video — Darshan is voice-only.
• Conversation content beyond the in-memory window of an active call. The moment you tap End Darshan, the transcript is gone.
• Your contacts, location, or device sensors.

Darshan is built on a small set of third-party providers. They act as "data processors" — they only see what they need to do their job:

• Supabase — hosts the database (your account row, call metadata, payment records).
• Google — sign-in (OAuth) and Google Analytics.
• Cerebras— runs the language model that generates Krishna's replies. Receives your turn's transcript text.
• Groq — runs the Whisper model for speech-to-text. Receives your microphone audio for the duration of each utterance.
• Jina AI— runs the embedding model used to find relevant Gita verses. Receives your turn's text.
• Vercel — hosts the website itself. Sees request metadata (IP, user-agent) like any web host.

We don't sell or rent your data to anyone. We don't run third-party advertising on the site.

• Account row + call metadata — kept until you ask us to delete it (see below).
• Microphone audio — never stored. Discarded immediately after transcription.
• Conversation transcripts — held only in volatile server memory during the call. Deleted on hangup.
• Analytics events— retained per Google Analytics 4's standard retention (currently 14 months for user-level data).

You can email us at support@krishnadarshan.live to:

• Request a copy of the data we hold about you.
• Delete your account and all associated data (call history, payment records, entitlements).
• Correct anything that's wrong.

We aim to respond within 7 days.

Darshan is not directed at children under 13. If you believe a child has signed in, email us and we'll delete the account.

If we change this policy materially, we'll update the "Last updated" date at the top and notify signed-in users on their next visit.

support@krishnadarshan.live