SlyshuPishu: it listens, it types

A coworker sent a seven-minute voice note, a lecture sits on your recorder and an interview needs to become an article. Listening again takes forever. SlyshuPishu, Russian for "I hear, I write", transcribes audio to text right in Telegram: send a recording and get a finished file. 🎧

About 840 people use it monthly. Recognition is based on the Whisper model, and a language model handles fixes and translation.

What you can send

• Voice messages from any chat, just forward them.
• Audio files from recorders, calls or podcasts.
• Video up to 400 MB and 150 minutes.
• YouTube links when you need a video's text.

Transcripts arrive as a TXT file, and an hour of audio takes about two minutes on average. ⚡

More than transcription

• Summaries when you need the gist, not every word.
• Error fixes for spelling and punctuation.
• Translation into other languages.
• Languages: Russian and English.

Price

The first 30 minutes are free. After that, minute packs according to reviews:

• 60 minutes for 129 rubles.
• 120 minutes for 199 rubles.
• 300 minutes for 399 rubles, about 80 per hour.

For comparison, online services usually charge several rubles per minute, and human transcription costs dozens of times more. Check your balance with /balance and buy packs via /pay. 💳

Getting clean text

1. Record without background noise. Cafés and streets hurt quality.
2. Speak close to the mic and avoid talking over each other in conversations.
3. Check names and terms. Rare surnames and niche vocabulary are harder.
4. Use summaries for long meetings and keep full transcripts for quotes.

Where it saves hours

A student records a 90-minute lecture and minutes later has searchable text for notes. A journalist transcribes an interview and pulls quotes without replaying it ten times. A manager gets a meeting summary and shares it with the team.

Even everyday voice notes are easier to read on the subway, in a meeting or late at night when sound is inconvenient. Forward the message and the text appears in seconds.

Privacy

According to the service, files are deleted automatically after transcription, and the privacy policy is available via /privacy. Still, avoid sending recordings with passwords, card details or others' personal data. 🔒

Who it helps

• Students recording lectures.
• Journalists and writers working with interviews.
• People who dislike voice notes and prefer reading.
• People with hearing impairments who find text easier.

Project news and a user chat are listed in the bot description, and support is available via /support. For important documents, always reread the full transcript, since AI can confuse similar-sounding words or numbers.