Any recording to a finished transcript
Transcript Generator for Audio and Video
Upload any recording, such as an interview, meeting, class or podcast, and get a transcript you can hand to someone: speaker names, a timestamp on every paragraph, and the choice of clean or word-for-word text. Most files are ready in about a minute. Download it as TXT for free, or as Word, PDF, subtitles or an NVivo / MAXQDA file on a paid plan.
Drop the recording you need a transcript of
Drop audio or video here
Audio or video, up to 5 hours. Or record now.
MP3, MP4, M4A, WAV, MOV, WEBM, OGG · up to 5 hours
We never use your files to train models. Files delete after 30 days, or right away when you choose.
A transcript, not a wall of text
Paragraphs by speaker, each starting with a timestamp. Speaker 1 becomes the real name everywhere after one edit.
Clean or verbatim
Clean reads like an article, without um, uh and stutters. Verbatim keeps every word, for legal review or conversation analysis.
The format the reader needs
Word for editors, PDF for sharing, SRT for video, NVivo and MAXQDA files for researchers.
How to generate a transcript
- 1
Upload or record
Drop an audio or video file, or record in the browser. Add names and terms you expect to hear.
- 2
Get the first draft
Text appears within seconds and speaker labels follow. Long files are emailed when done.
- 3
Name speakers and fix errors
Rename speakers, move a line to the right person, merge split sentences and correct words while listening.
- 4
Download
Pick the format and the clean or verbatim version, then export. Keep it in My transcripts or delete it right away.
What separates a usable transcript from a rough one
Speaker names, not Speaker 1
A transcript with generic labels gets sent back. Rename each speaker once before export; it takes seconds and every line updates.
Timestamps people can check
Quotes get questioned. A time at the start of each paragraph lets anyone find the moment in the recording; in our editor, clicking a line plays it.
One pass by ear
Even good machine transcripts get some names, numbers and crosstalk wrong. Listen through the tricky parts at 1.5× or 2× and fix them in place.
Search before you scroll
Looking for the part where someone mentioned a figure or a name? Search the transcript and jump straight to it.
Transcript formats and what each one contains
| Format | Speakers | Timestamps | Plan | Use it for |
|---|---|---|---|---|
| TXT | Yes | Every paragraph | Free | Notes, pasting into other tools |
| DOCX (Word) | Yes, in bold | Every paragraph | Paid | Editing, sharing, quoting |
| Yes | Every paragraph | Paid | Sending a final copy | |
| SRT / VTT | VTT only | Every cue | Paid | Captions for video |
| NVivo format | Yes, "[hh:mm:ss] Name:" | Every paragraph | Paid | Auto-coding by speaker in NVivo |
| MAXQDA format | Yes | "#hh:mm:ss-0#" per paragraph | Paid | Linking text to audio in MAXQDA |
All exports use your edited text and speaker names.
Questions
Is the transcript generator free?
Yes, for 3 files a day up to 30 minutes each and 120 minutes a month, with TXT export. Paid plans start at $9.99 for 10 hours, and those hours never expire.
Can it summarize the transcript?
No. We make the full transcript, not summaries or action items. You can search it, edit it and export it.
Does someone type my transcript by hand?
No. Speech recognition models make the transcript and you edit it in the browser. There is no human transcription service.
What happens to my recording afterwards?
It is used only for your transcript, never for training models. Audio and text delete after 30 days, after 24 hours if you choose, or straight away when you press Delete.
Related
Transcribe your first file now
Free: 3 files a day, first 30 minutes of each. Sign in with Google to start.
Choose a file