Time coded transcript generator
A time coded transcript is a written record of a recording in which each passage starts with the timecode where it is spoken, so you can find any line in the audio or the edit timeline.
Recent files, saved in this browser
Keep this tab open. The transcript appears below as it comes in.
Choose the same file again to play it from any timecode.
The first 10 minutes of up to 3 files a day free, no sign-up Then 10 hours for $9.99, paid once Audio is transcribed, not kept Unused minutes refunded within 14 days
How it works
- 01
Add your recording
Drop an audio or video file on the box at the top of the page. MP3, WAV, M4A, MP4, MOV and most other formats work, up to 3 hours.
- 02
Transcribe it
The first 10 minutes of up to 3 files a day are free. For the whole file, use minutes from a pack. The transcript appears as it comes in, with a timecode on every sentence.
- 03
Set the timecodes and download
Pick the timecode format, frame rate, start timecode and how often a timecode appears, then download TXT, SRT, VTT, CSV, DOCX or JSON. Changing a setting never transcribes again.
What a time coded transcript looks like
A time coded transcript pairs every passage of text with the moment it is spoken. The same recording can be written with different timecodes, depending on who reads it.
[HH:MM:SS] for notes and reports
The common format for interviews, research and legal transcripts. Hours, minutes and seconds in square brackets at the start of each passage:
[00:00:00] The lighthouse at Cape Wrath was built in 1828.
[00:00:15] Didn't the keepers get lonely?
[00:00:28] Winter storms cut them off for weeks at a time.HH:MM:SS:FF for video editing
SMPTE timecode adds frames after the seconds, so a passage can be found on the exact frame of an edit timeline. The number of frames per second depends on the video: 23.976 or 24 for film, 25 in Europe, 29.97 or 30 in North America. At 29.97 fps, drop-frame timecode (written with a semicolon, 01:00:00;00) keeps the timecode in step with the clock.
01:00:00;11 The lighthouse at Cape Wrath was built in 1828.
01:00:37;04 In January 1862, a supply boat finally reached them.SRT subtitles
An SRT file numbers each subtitle and gives its start and end time, with a comma before the milliseconds. YouTube, Premiere Pro, DaVinci Resolve and Final Cut Pro all import it.
1
00:00:00,400 --> 00:00:03,400
The lighthouse at Cape
Wrath was built in 1828.
2
00:00:15,300 --> 00:00:17,200
Didn't the keepers get lonely?WebVTT subtitles
WebVTT is the subtitle format of web video players. It starts with a WEBVTT line and uses a full stop before the milliseconds:
WEBVTT
00:00:00.400 --> 00:00:03.400
The lighthouse at Cape
Wrath was built in 1828.Timecodes that match your edit
A camera file or a sequence rarely starts at 00:00:00:00. Set the frame rate and the start timecode, and every timecode in the transcript, the SRT and the VTT is shifted to match. Change the setting at any time: the transcript is relabelled on the spot, not transcribed again.
A timecode every sentence, paragraph or 30 seconds
Editors usually want a timecode on every sentence. Transcription services often put one every 30 or 60 seconds, or at each new paragraph, so the text reads more easily. Choose every sentence, every paragraph, or every 10, 15, 30 or 60 seconds; the timecode always lands at the start of a sentence, never in the middle of a word.
Long recordings
Files up to 3 hours work: lectures, podcasts, audiobook chapters, full interviews. The file is decoded in your browser and sent in short parts, so a long upload does not start over when one part fails. A part that fails is retried, and if it still fails, its minutes are not charged.
Free, then pay once
The first 10 minutes of up to 3 files a day are free, with timecodes on every sentence and TXT and SRT downloads. Minute packs are paid once and never renew, and each can be used for 12 months; they unlock whole files, SMPTE frames, the start timecode, interval options and all six downloads. See pricing.
Questions
What is a time coded transcript?
A transcript in which each passage starts with the timecode where it is spoken, such as [00:12:05] or 01:00:12:05. It lets you find any line in the recording or on an edit timeline without listening through it.
Is it free?
The first 10 minutes of up to 3 files a day are free, with no sign-up. The free result has a timecode on every sentence and downloads as TXT or SRT. For longer files and the editing formats, buy a minute pack once; there is no subscription.
Which timecode formats can I use?
[HH:MM:SS] for documents, HH:MM:SS.mmm with milliseconds, and SMPTE HH:MM:SS:FF frames at 23.976, 24, 25, 29.97 (drop-frame or non-drop) or 30 fps. You can also add a start timecode such as 01:00:00:00.
How do I make the timecodes match my Premiere Pro or DaVinci Resolve timeline?
Choose HH:MM:SS:FF, set the frame rate of your sequence, and enter the timecode of the first frame of your clip as the start timecode (often 01:00:00:00 for a sequence, or the source timecode of a camera file). Every timecode in the transcript and the downloads then lines up with the timeline.
Where does each timecode come from?
The speech model times every word. A timecode marks the start of the first word of its passage, so it points at where that passage begins in the audio. These timings are estimates and can be off by a fraction of a second.
How accurate is it?
Every automatic transcript has mistakes, and the model can occasionally add words that were not said, especially during silence or music. Check names, numbers and anything you will quote against the recording. For court, medical or other records where accuracy matters, have a person review the transcript.
Which languages work?
The language is detected automatically. The speech model (Whisper large-v3-turbo) handles about 100 languages; English and other widely spoken languages work best, and clear audio always helps.
Do you keep my audio or transcript?
Your file is cut into short pieces in your browser, and each piece is sent to be transcribed and then discarded. We do not keep the audio or the text on our servers. The transcript is saved only in this browser on your device, and you can delete it at any time.
Can I get my money back?
Yes. Unused minutes are refunded in full within 14 days of purchase, and minutes for parts that could not be transcribed are never charged. See the refund policy.
Can it label speakers?
Speaker names are not added automatically yet. You can click any line and type a name, such as "Interviewer:", before you download.