Forward a voice message to the bot and the text comes back in a couple of seconds. Five minutes of audio takes about one. The first 20 minutes every month are free.
How it works
Send a voice message in a DM — or add the bot to a group, where it handles voice messages on its own. Molnia listens to the recording and writes back in first person: the key point on the first line, then each topic as its own paragraph.
Voice messages, audio, video notes, video, and files: mp3, m4a, wav, ogg, mp4. Forwarded from another chat works too. One recording — up to 3 hours.
Recognizes speech regardless of the recording's language and writes to the point.
The key point sits on top, then each topic gets its own paragraph. The longer the recording, the more detail comes back.
Five minutes of audio comes back in about a second. An hour of audio takes a few seconds — faster than Telegram's own built-in transcription.
Just an ordinary Telegram bot: no apps, no signup, no access to your other chats. The temporary copy is deleted right after processing, and the transcript and summary never get written to our database.
Add the bot to the chat — it transcribes voice messages on its own, right under each one. Nothing to set up.
Pricing
20 minutesa month, in DMs
About 10–15 typical voice messages. Enough to test it on your own recordings and see if you need more.
Start free1,000 minutesa month, $5 by card
Over 16 hours of audio — more than enough for a month of voice messages. Pay by card; renews automatically, cancel anytime.
Open in TelegramIn group chats, the first 100 minutes a month are free for everyone — a subscription is only needed beyond that.
Forward the voice message to the bot @molniavoicebot in a DM. In a group, just add the bot — it transcribes voice messages on its own. You'll get back a text: the key point on the first line, topics below as separate paragraphs. Five minutes of audio comes back in about a second.
Telegram Premium's built-in transcription needs a subscription and gives you a word-for-word transcript as long as the recording itself. Molnia works without Premium and compresses the voice message down to what matters: the key point on top, topics below. An hour of audio reads in about half a minute. A detailed comparison with Premium is on its own page.
Yes. Molnia transcribes voice messages, audio, video notes, video messages, and files forwarded from other chats. One recording — up to 3 hours.
Formats: Telegram voice messages, video notes, mp3, m4a, wav, ogg, mp4. The recording's language is detected automatically — Russian, English, and others — and the summary comes back in that same language.
Free — 20 minutes of transcription a month in DMs, no signup needed. Beyond that, a subscription: 1,000 minutes a month for $5 by card; renews automatically, cancel anytime. In group chats, the first 100 minutes a month are free for everyone.
Yes. Add the bot to the chat and it transcribes voice messages and video notes on its own, right under each one, for everyone to read. Audio files and forwarded recordings are left alone — reply to those with /sum. Chat settings and the off switch live in /settings.
The bot doesn't have access to your other chats. It makes a temporary copy of the recording and deletes it after processing; the transcript and summary aren't saved to our database, but the original message and the reply stay in Telegram.
No. Open @molniavoicebot in Telegram, send a voice message, and you'll get the text right away. Nothing to install, nothing to register.
The reply lands before you'd even manage to switch to another chat.
Send a voice message