Free tool
Audio to Text: Transcribe Audio and Voice Recordings
Upload an audio or video file and get a text transcript with timestamps and speaker names — a speaker is named when they say their name aloud, and gets a letter when no name is said. Copy it or download TXT, Word, SRT or VTT. Audio is deleted as soon as the transcript is ready.
What you get
A transcript split into short lines, each with the time it was spoken and the speaker who said it. After the speech is turned into text, the voices are told apart, and a speaker is named where their name is said aloud; a speaker whose name never comes up gets a letter: A, B and so on. Click a speaker to rename it, and every export uses the new name.
Above the transcript you can search for a word, show only one speaker, or jump to a timestamp. The Copy button puts the whole text on your clipboard with speaker names. Download gives you the same text as plain text, text with timestamps, SRT or VTT subtitles you can load into a video player, a Word document, Markdown for Notion or Obsidian, or a PDF.
Which formats work
Most recordings can be uploaded exactly as they come off the device — there is no need to convert them to MP3 first. The engine decodes the file on the server, video included, and only the sound is transcribed.
| Where the recording comes from | Usual file type |
|---|---|
| iPhone Voice Memos | .m4a |
| WhatsApp voice message | .opus |
| Telegram voice message | .ogg |
| Zoom local recording | .m4a (audio only) or .mp4 (with video) |
| Dictaphone or field recorder | .wav or .mp3 |
| Screen recording or phone video | .mp4 or .mov |
The upload limit is 350 MB. Speech-only audio is usually well within it; video is not — a few minutes of high-resolution phone video can already be too large. For a long video, open it in MeetWave instead: the app pulls the audio out of the video in your browser before uploading.
MP3 to text and WAV to text: file size and length
MP3 and WAV are both accepted as they are. What differs is size: MP3 is compressed, WAV is not, so the same minute of speech weighs about ten times more as WAV. One minute is 60 seconds, and the limit is 367,001,600 bytes (350 MB of 1,048,576 bytes), so the minutes that fit are the limit divided by the size of one minute.
| Format | Bytes per second | Size of one minute | Fits in 350 MB |
|---|---|---|---|
| MP3, 128 kbps | 128,000 bit/s ÷ 8 = 16,000 B/s | 0.9 MB | 382 min (6.4 h) |
| WAV, 44.1 kHz, 16-bit, mono | 44,100 × 2 B × 1 = 88,200 B/s | 5.0 MB | 69 min |
| WAV, 44.1 kHz, 16-bit, stereo | 44,100 × 2 B × 2 = 176,400 B/s | 10.1 MB | 34 min |
| WAV, 48 kHz, 24-bit, stereo | 48,000 × 3 B × 2 = 288,000 B/s | 16.5 MB | 21 min |
A stereo WAV at CD quality therefore fills the limit in about half an hour, well short of the first 60 minutes that are free. If a WAV is too large, save it as mono or as an MP3 rather than cutting it. Do not convert a WAV to MP3 before uploading when it already fits: the extra step buys nothing. A phone call recorded on one channel puts both voices on one track, and they are still told apart as two speakers.
What happens to your file
You do not need an account, so we keep as little as possible and for as short a time as possible. The transcript is not used to train any model, and nothing about its content goes into our logs or analytics.
| What | Where it is kept | How long |
|---|---|---|
| Your audio or video file | Encrypted storage used only for this job | Until the transcript is ready, then deleted — every stored version of it |
| The working copy on the transcription server | The server that runs the speech-to-text engine | Deleted together with the file |
| The transcript | Encrypted storage, under a random job id | Deleted within 48 hours, or at once when you click "Delete now" |
| The file name | Only in your browser, so a reload can show it | Never sent to us |
Getting a clean transcript
Most of the accuracy is decided when the audio is recorded, not when it is transcribed. Three things matter most. The distance to the speaker: a phone or microphone within arm's length gives clear words, one across the room gives gaps. Background noise: music, traffic and a humming air conditioner compete with speech. Crosstalk: when two people talk at once, both lines suffer and the speaker labels can merge.
If you record the file yourself, a few seconds of test recording played back before the real thing catches most problems. For a fuller walkthrough — from recording to editing the text — read the step-by-step guide to transcribing audio, and if you need to choose a paid service for regular work, compare transcription software before you buy.
Transcript vs meeting summary
A transcript is every word in order: right for quotes, interviews, subtitles and records. A summary is what the conversation decided — the points, decisions and action items — and for a one-hour meeting it is the difference between reading ten pages and reading one. This page gives you the transcript of a file you already have: the first 60 minutes free, then one credit per started 30 minutes. MeetWave meeting summaries start earlier: the app records your Zoom, Google Meet or Teams call without a bot joining it, and turns it into a summary with action items when the call ends.
How to convert audio to text
- Choose your fileDrag an audio or video file onto the page or pick it from your device: MP3, WAV, M4A, OGG, OPUS, FLAC, AAC, WMA, WEBM, MP4, MOV, MKV or AVI, up to 350 MB. The page reads its length before anything is uploaded.
- Start the transcriptionClick Transcribe. The file goes straight to encrypted storage, and the speech-to-text engine starts. A timer shows how long you have been waiting; you can reload the page and the result will still arrive.
- Read, search and renameThe transcript appears line by line with timestamps and speakers: a name where it was said aloud, a letter such as A otherwise. Search it, show one speaker at a time, and click a speaker to rename it.
- Copy or download — then deleteCopy the text or download TXT, SRT, VTT, Word, Markdown or PDF. The audio is already gone; the transcript is deleted within 48 hours, or delete it now with one click.
Frequently Asked Questions
Yes, the first file is free with no account: up to 60 minutes of it are transcribed, and a longer file gets its first 60 minutes with a note saying so. After that, sign in: new accounts get 30 days of Pro, which includes the tools, plus 3 credits that stay after it; on the free plan a file costs one credit per started 30 minutes — a 45-minute recording is 2 credits. There is also a daily limit on free transcriptions for everyone together; when it is reached, the page offers the sign-in instead.
It is uploaded to encrypted storage used only for this job, transcribed, and deleted as soon as the transcript is ready — including the working copy on the transcription server. The transcript is deleted within 48 hours; "Delete now" removes it immediately. The file name stays in your browser, and neither the audio nor the text is used to train models.
The engine detects the spoken language by itself — there is nothing to choose. Quality is best in widely spoken languages and on clear recordings; the transcript comes out in the language that was spoken, not translated.
Yes. After transcription the voices are told apart, and each speaker is named when their name is said aloud — "Thanks, Anna" or "This is Tom". A speaker whose name is never said gets a letter: A, B and so on. Click a speaker to rename it — the new name is used in every download. Two people talking over each other, or very similar voices, can end up as one speaker.
It depends on the length of the file and on whether the transcription engine has to start up first — the first run after a quiet spell takes longer. While you wait, the page shows the elapsed time and an estimate based on the length of your file. You can reload the page or come back to it later the same day — the result will be there.
Yes. The same transcription is available as JSON through the MeetWave speech-to-text API: text with timestamps, and speakers named when they say their names aloud. It costs $0.20 an hour of audio with speaker names, and the first hour is free.
This tool gives you the transcript of one file you already have. MeetWave records your Zoom, Google Meet and Teams calls without a bot joining, then turns them into a summary with decisions and action items. You can open the same file in MeetWave to get that summary — the free plan includes 600 minutes of recordings a month.
