The first five minutes of any file are free: readable, searchable, copyable and synced to the audio, three files a day. Finishing the whole file, the exports and the notes take one credit. There is no subscription and no account.

Which languages does it handle?

The language is detected automatically. English, Turkish, Spanish and Portuguese are the ones this site runs in; the model behind it transcribes dozens more. If detection picks the wrong one, you can set the language and run again.

Does it tell speakers apart?

Not yet. Paragraphs break where the speech pauses, which usually lines up with a change of speaker, but no names or labels are added. It is on the list.

What happens to my recording?

It is decoded in your browser and sent to the speech model as plain audio in short pieces, then discarded. We do not keep the audio. The transcript is saved to your history so you can reopen or share it; you can delete it from your account.

How long does it take?

The first sentences arrive in a few seconds. A one hour recording takes a few minutes to finish, and the text fills in as it goes, so you can start reading immediately.

What files can I upload?

Almost any audio or video: MP3, M4A, WAV, AAC, OGG, FLAC, MP4, MOV, WEBM, MKV and more, up to 600 MB and three hours. For a video, only the sound is used.

Yes. Double click any sentence to edit it. Every export, including the PDF and the caption files, reflects your corrections.

Do the caption files work on YouTube?

The SRT export is the format YouTube, LinkedIn, Vimeo and every editor accept. Cues are kept to two lines of forty two characters and no longer than six seconds. For short social video, the Subtitle Generator on this site previews the cues on the video itself.

Audio Transcription

Any recording or video into text you can read, search and send. The first five minutes are free and follow the audio as it plays. One credit finishes the file, the exports and the notes.

No accountOne payment per file, no subscriptionThe audio is not kept. Only the text stays in your history.

Example: a voice memo after a client call. This is what came back for it, word for word. Press play and follow along.

What one credit gives you

Full transcript, every minute
Speaker labels: who said what
TXT, DOCX and PDF
SRT and VTT captions
Notes: summary, key points, action items, chapters

Is the video on YouTube, TikTok or Instagram? Paste the link instead

Drop in a voice memo, an interview, a meeting recording, a podcast episode, a lecture or a video. The audio is decoded in your browser and sent as plain sound in short pieces, so the file itself is never uploaded and the first sentences come back before the recording has finished playing once. Every paragraph carries the moment it was said: click it to hear that moment, and the words light up as the audio runs, so you can check any line against the recording without scrubbing.

The first five minutes are free, fully readable, copyable and playable. One credit finishes the file, turns it into TXT, DOCX, PDF and caption files, and writes working notes: a summary, the key points, the action items and chapters with timestamps. No account, no subscription. The audio is not kept; only the text stays in your history so you can reopen or share it.

Every line knows when it was said

A transcript page floating on a warm field. A paragraph is highlighted in orange with the timestamp 0:31 beside it, and a waveform above shows the playhead at the same point. — Click a paragraph and the recording jumps there. Press play and the words light up as they are spoken, so a doubtful name or number takes two seconds to check.

How it works

Drop a recording or a video
MP3, M4A, WAV, MP4, MOV, WEBM and most other formats, up to three hours. The audio is decoded in your browser and only the sound goes out, in short pieces.
Read while it listens
The first sentences arrive in a few seconds. The transcript fills in as the recording plays through, with the language detected on the way.
Check it against the audio
Every paragraph carries its timestamp. Click to hear that moment; double click to correct a word. Search finds any phrase.
Take the files and the notes
One credit finishes the whole file and exports TXT, DOCX, PDF, SRT and VTT, plus notes: summary, key points, action items and chapters.

Notes written from your own words

A notes sheet floating over a transcript: a short summary, three key points, two action items with checkboxes, and four chapters with timestamps. — Summary, key points, action items and chapters, every one taken from the transcript. For a podcast or a video it also drafts the episode title and description.

What changes when the tool is one file, not a plan

Here

Before the first result: Nothing to sign up for
What is free: First five minutes, readable and playable
What you pay for: One file, one payment
Exports: TXT, DOCX, PDF, SRT, VTT
Notes: Summary, key points, action items, chapters
Your audio: Not kept

Meeting note apps

Before the first result: Account, often a calendar link
What is free: A monthly minute cap, three minute files on some
What you pay for: A monthly subscription
Exports: Usually on paid plans only
Notes: Yes, on paid plans
Your audio: Stored in your workspace

What accuracy honestly depends on
Clear speech in a quiet room comes back nearly perfect in the languages this page runs in, and in dozens of others. What lowers accuracy is the same for every transcription service: two people talking at once, a phone in a bag, music under the voice, and names or product terms the model has never seen. That is why the transcript sits next to the audio here rather than in a separate download: the places worth checking are the places you can hear. Speaker names are not detected yet; paragraphs break where the speech pauses.

Before you record, if you can

Put the phone on the table, not in a pocket.
Cloth over the microphone is the most common reason for missing words.
One voice at a time.
Interruptions are transcribed as they happen, which is rarely what anyone meant.
Say names and numbers slowly once.
The transcript keeps the timestamp, so checking them later takes seconds.
Turn off music under the voice if you can.
Lyrics compete with speech.
Long file?
Three hours per credit is the limit. Split a longer lecture into parts.

Questions

Unlock the rest of this tool

Subscribe for unlimited use, or buy credits and pay only for what you need.

Best value

5 transcripts

$9.99

5 files, up to 3 hours each
Every export, the notes
No expiry
$2 each

Popular

1 transcript · single file

$2.99

The whole file, up to 3 hours
TXT, DOCX, PDF, SRT, VTT
Notes: summary, key points, action items

Cancel anytime Easy refund, just write us Secure checkout · Polar