How to turn an iPhone Voice Memos recording into text

Potok transcribes Russian speech; this site has been translated into English for convenience.

An iPhone Voice Memos recording in m4a format is transcribed into text through the Potok Telegram bot @potok_now_bot: send the file and the bot returns finished text with punctuation placed. For a larger file the bot's mini app fits. 60 minutes of transcription per month are free; for dictating your own text by voice on a computer there is the separate Potok app.

A recording from the built-in Voice Memos app on iPhone is saved in m4a format, an ordinary audio file that can be forwarded to a messenger like any other file. Such a recording can be transcribed straight from the phone, without a computer and without installing programs.

Transcribing an iPhone Voice Memos recording into text through the bot

1. Find the bot @potok_now_bot in Telegram and start it with the /start command.
2. Send the m4a recording file as a document: this way it arrives without compression, however long the conversation or note is.
3. Wait for the reply: a short recording comes as text right in the chat, a long one as a separate file; the format for long replies (TXT or DOCX) is chosen in advance in the settings.
4. If the recording has several voices and you need speaker labels or timecodes, turn them on in the bot settings (/settings) before sending the file: with several voices, transcription is harder than with one, and why is covered in the article why transcribing meetings is harder than ordinary dictation.
5. A larger file is easier to upload to the bot's mini app: sign in by email or through Telegram, a limit of 2 GB per file.
6. All transcripts are saved in the My recordings section, and you can open them later both in the bot and in a browser.

What comes back for an m4a file

• A short transcript comes as text right in the chat; a long one as a file in the chosen format, TXT or DOCX.
• The text already has punctuation and capital letters in place, and the bot also corrects obvious recognition errors. It will not shorten or retell the recording, unless you press the “Summary” button yourself.
• Under the transcript there are buttons: “Summary” (the main points and tasks by speaker), “Tasks and deadlines”, “Ask” (your own question about the content of the recording) and “Other format”, an export to TXT, DOCX, Markdown, PDF, SRT or VTT.
• If you turn on timecodes and speaker labels in advance, for a recording with several voices this also appears in the text, under the labels “Speaker 1”, “Speaker 2”; the bot does not know the participants' names.
• The finished transcript stays in the My recordings section: you can open it again at any moment, from the bot or from the web version.

When it is better not to trust the bot with a voice recorder recording

If the recording was made in a language other than Russian or a translation is needed, the bot will not help: it works only with Russian speech. It also does not hear a live conversation: it needs an already finished file, and it does not accept a broadcast during a call or meeting. The Potok bot is cloud-based: the audio goes to servers in Russia, and without the internet transcription will not start. It tells the voices in a recording apart only by the numbers “Speaker 1”, “Speaker 2”, and the participants' names have to be added by hand. Ready-made transcription services can be compared with each other in a separate review, the comparison of transcription services.

Frequently asked questions

Primary sources: speech2text.ru · audio-2-text.ru · turboscribe.ai · turboscribe.ai