Record anything
Meetings, lectures, calls and voice notes, with the screen off. Pause and stop from the notification, let it pause itself during a phone call, and it saves in parts as it goes — so a crash doesn't cost you the recording.
Record a meeting, a lecture or a call, and get back a clear summary: what was decided, who does what by when, and answers to anything you ask about it. Run the AI on your phone, on your own API key, or on Ganakya Premium.
Press record and stay in the conversation. When it ends, Saaransh writes the transcript, then a summary that leads with what matters: the decisions, and the tasks with an owner and a due date.
It keeps going with the app closed, and a notification says when the summary is ready. Every task lands in one list across all your recordings.
Meetings switch between Hindi and English mid-sentence. Saaransh writes the summary in the language people mostly spoke — Hinglish stays Hinglish — or in the one you choose.
Everything in this list works in the build being tested on Google Play now.
Meetings, lectures, calls and voice notes, with the screen off. Pause and stop from the notification, let it pause itself during a phone call, and it saves in parts as it goes — so a crash doesn't cost you the recording.
Every line timestamped, in the language people actually spoke — Hindi, English, Hinglish and more. Import any audio file, or share one to Saaransh from another app.
A title, the overview, decisions, action items with who and when, topics with where they start, and the speakers. Short, detailed, bullet points or meeting minutes.
The action items from all your recordings in one list — due today, this week, later or with no date — each linked to the recording it came from.
“What did we decide about the price?” Answers come from the transcript, with the times they rest on.
On this phone, offline after a one-time download. With your own OpenAI, Gemini or compatible API key. Or Ganakya Premium on our server — nothing to set up.
Audio, transcripts and summaries are stored on the phone, not on a server. Audio leaves it only for the AI engine you pick.
Sign in with your phone number — no email or password. Premium follows your number, and the account is open on one phone at a time.
Whisper writes the transcript and Qwen 2.5 the summary, both running on the phone after a one-time download of about 2 GB. Works offline; nothing leaves the phone.
OpenAI, Gemini, or any OpenAI-compatible service such as Groq or OpenRouter. Audio goes straight from your phone to that provider, and the key stays encrypted on the phone.
Nothing to set up: Google's Gemini on our server writes the transcript and summary, and each piece of audio is deleted as soon as it's transcribed. A Google Play subscription with a 7-day free trial.
Transcripts and summaries carry on with the app closed and pick up after a restart. A failed summary keeps its transcript, so trying again starts where it stopped.
A recording of a meeting is about as private as data gets. Saaransh keeps your audio, transcripts and summaries on your phone, and our server holds only your account and Premium — never what you recorded.
Saaransh is being tested on Google Play ahead of its public release. When it launches, the download button on this page goes live.