Free Tool

Transcribe a Voice Recording

Upload a voice memo from your phone or a voice message saved from WhatsApp or Telegram, and get the text back with timestamps and a label for each speaker. The first recording is free up to 30 minutes, with no account. The audio is deleted as soon as the transcript is ready, and the text can be copied or downloaded as TXT, Word, SRT or VTT.

First file free up to 30 min · then 1 credit per 30 min (from $0.06) · New accounts get 3 credits

Any other audio or video file? Use the audio to text converter — same engine, same limits.

Voice recording to transcript: an example

Input: a 2-minute voice memo recorded on an iPhone after a call — two people, one phone on the table (New Recording 14.m4a).

Output:

TimeSpeakerText
00:00Speaker 1Okay, it's recording. Let's go over what we agreed with the landlord.
00:05Speaker 2The heating gets fixed by Friday, and they send someone Thursday afternoon.
00:12Speaker 1And the rent for October stays the same?
00:15Speaker 2Same, but November goes up. They'll email the new amount.
00:21Speaker 1Then I'll forward that email to you as soon as it comes.
00:25Speaker 2Good. And ask about the parking spot again.

Rename "Speaker 1" and "Speaker 2" to real names on the page, and the names carry into every download.

Export a voice memo from iPhone

Open Voice Memos, tap the recording, then tap the three-dot button and choose Share. Pick "Save to Files" to keep it on the phone, or AirDrop it to a Mac. The file arrives as an .m4a — upload it here unchanged. If you are on the same iPhone, you can also open this page in Safari and choose the recording from Files directly in the upload dialog.

The name of the memo stays in your browser. It is shown on the page so you know which file you are looking at, and it is never sent to our servers — the upload is stored under a random id.

Save a WhatsApp or Telegram voice message as a file

A voice message lives inside the chat until you save it. In WhatsApp on an iPhone, press and hold the message, choose Forward, then the share icon, and save it to Files. On Android, WhatsApp keeps voice messages in its media folder on the phone, where a file manager can reach them, or you can share the message to Google Drive. In Telegram Desktop, right-click the voice message and choose Save As.

Messenger voice notes are often recorded close to the mouth, which is good for transcription — but they are also often recorded while walking. If the text has gaps, the wind or the street was louder than the voice at that moment.

Android recorder files

Recorder apps on Android save to the phone's storage, usually as .m4a or .mp3, sometimes as .wav when a high-quality mode is on. Open the recording in the app, use Share to send it to Drive or to your computer, or find it in a file manager under a folder named after the app. All three formats are accepted here as they are.

A few older phones record calls and memos as .amr or .3gp. Those are not on the list above; convert the file to MP3 or open it in MeetWave instead.

Why a voice note from a messenger may be .opus — and why that's fine here

Opus is an audio codec built for speech over the internet: it keeps a voice clear at a very small file size, which is why messengers use it. Many desktop players and older editors do not open .opus files, and that leads people to convert them to MP3 before transcribing. There is no need to here — the voice recording transcription runs on a server that decodes Opus directly, so converting would only add a step and lose a little quality.

Your browser may not be able to play the file, so the page may not show its length before the upload. That is expected: the length appears with the transcript.

From voice recording to transcript — and what happens next

The first voice recording is free up to 30 minutes with no account; signing up gives 3 free credits, one per started 30 minutes of a file. If you would rather do it yourself on the phone, the three ways compared include the transcription built into your phone.

WhatWhere it is keptHow long
Your audio or video fileEncrypted storage used only for this jobUntil the transcript is ready, then deleted — every stored version of it
The working copy on the transcription serverThe server that runs the speech-to-text engineDeleted together with the file
The transcriptEncrypted storage, under a random job id24 hours, or until you click "Delete now"; a storage rule removes anything left within 48 hours
The file nameOnly in your browser, so a reload can show itNever sent to us

How to transcribe a voice recording

Step 1

Get the recording out of the app

Share the voice memo or voice message to Files, Drive or your computer, so it becomes an ordinary file — the sections below show where each app puts it.

Step 2

Drop it on this page

M4A from Voice Memos, OPUS or OGG from WhatsApp and Telegram, MP3 or WAV from recorder apps — up to 350 MB. Click Transcribe.

Step 3

Read the transcript

Each line has a timestamp and a speaker label. Rename the speakers, search for a word, then copy the text or download it.

Frequently Asked Questions

Yes. Save the voice message as a file first — on Android it is already in the WhatsApp media folder, on an iPhone you share it and save it to Files — and upload that file here. WhatsApp voice messages are .opus files, which this tool accepts directly: you do not need to convert them to MP3.

Without an account you get the first 30 minutes, marked as partial. Sign up and you receive 3 free credits; the whole file then costs one credit per started 30 minutes. For regular recordings MeetWave itself is the better fit: a file of up to three hours becomes a transcript and a summary, within the free plan's monthly minutes.

Usually, yes — what matters more than the phone is the distance to the speaker and the background noise. A phone held or placed within arm's length of the person talking gives clean text; a phone in a pocket, a recording next to a coffee machine or two people talking over each other gives gaps and mixed-up speakers. If you can, record a few seconds of a test first and listen back.

Engine: Whisper large-v3-turbo model card (Hugging Face) · WhisperX (GitHub) · pyannote speaker-diarization-3.1 model card