Live
Subtitle Generator
SRT and VTT subtitles from a video or a recording
Video to Text
LiveTranscribe video to text online: upload a lecture, meeting, interview, or clip and get a transcript to read, copy, and download, in 18 languages.
What it does
- Handles common phone, camera, and screen recording formats
- Auto language detection with optional timestamped sections
- Copy in the browser or download a plain text file
Your files
Gone within the hour.
Files are uploaded over HTTPS, used only to run this tool, and deleted from our server automatically about an hour later. Never sold, never used to train anything.
Speech is processed by our self-hosted faster-whisper worker. Your recording is not sent to a third-party AI service.
How to transcribe a video to text
Add one video
Drop it onto the box or click the box to browse. It takes one video at a time, up to 50 MB and 60 minutes long. When your browser can read the length, it does so before anything is uploaded, and on a longer video it says so and keeps Transcribe video switched off rather than letting you wait for a refusal. Where your browser can play the format you get a preview; where it cannot, the page says the server will still check it.
Tell it what it will hear
Spoken language starts on Auto detect, and choosing one of the 18 languages helps with names and short clips. Include timestamps is on to begin with and puts a marker such as [00:01:23] at the start of each spoken section; untick it for one block of plain text. Names or special terms takes up to 400 characters of spellings you want used, such as a product or a person's name. Play a sound when ready is also on, so you can leave the tab.
Press Transcribe video
A bar shows the upload. If other videos are ahead of yours, the page says how many and keeps your place. Once it starts, it shows Preparing audio and then Transcribing, with an estimated time that is worked out from how fast this server is getting through your own file.
Copy it or download it
The transcript appears in a box you can read and select. Press Copy transcript to put all of it on your clipboard, or Download transcript.txt for a plain text file. Save it before you leave: the video and the transcript are deleted from our server within about an hour. Transcribe another starts again with a new file.
Frequently asked questions
How do I transcribe video to text online?
Upload one video, choose its language or leave Auto detect selected, then press Transcribe video. It runs online in your browser with no account and nothing to install. While the job runs the page reports the stage it has reached and an estimated time remaining, refined from the speed measured on your own file rather than a fixed guess, so a long recording tells you where it is instead of showing a spinner. If you would rather not sit and watch it, leave Play a sound when ready ticked, as it is from the start, and a short chime tells you the transcript has arrived. You can read and copy the result in the page or download it as transcript.txt.
Which video formats can I transcribe?
The converter accepts common MP4, MOV, M4V, WebM, MKV, AVI, MPEG, OGV, FLV, and transport stream files. The video must contain a readable audio track: one with no audio stream at all comes back as "This video has no audio track to transcribe", and a video that has audio but no speech in it comes back as a message saying so rather than as an empty transcript. Before anything is uploaded, your browser checks that the file is typed as a video or has one of those extensions, that it is under 50 MB, and, when it can read the length, that it runs no longer than 60 minutes. The rest is decided by the server, which reads the first four kilobytes: a file it cannot read as a video or a sound file is turned away at upload, and one too damaged to read is stopped when the worker first opens it, before any transcription is attempted.
What are the video size and length limits?
One video per job, up to 50 MB, and up to 60 minutes of running time. The length is read in the browser from the video itself, so a file that is too long is refused before the upload starts rather than after you have waited for it, and the server checks the real length again when it opens the file. That second check measures the longest track in the file rather than the picture, which matters if you convert video to text free from a recording that is mostly a still frame: a webinar or a podcast with a two second loop on screen counts as its full spoken length, not as two seconds. The current free allowance is five speech jobs a day per connection, and it is shared with Subtitle Generator because both tools run the same transcription pass on the same endpoint: five between the two, not five each. If the speech worker is already busy your video waits in line rather than failing, and the page shows its place in the queue and counts it down as the videos ahead of it finish. Only when ten videos are already waiting or running does the site ask you to come back a little later instead of adding an eleventh.
Can the transcript include timestamps?
Yes. Keep Include timestamps selected and every spoken section starts with a marker in [HH:MM:SS] form, counted from the start of the video, one section to a line. A marker is the start of a spoken section, not of a word: the transcript is not word by word timed, so you can see where a passage begins but not where a particular word falls inside it. Silence is dropped before the speech model ever hears it, which is why a recording that opens with two quiet minutes produces no empty lines and why the markers jump across a long gap instead of counting through it. Turn timestamps off and the same sections are joined into continuous plain text. The markers are written into the text file you download, so they survive being pasted somewhere else. If you want a subtitle file with the timings kept as cues, use Subtitle Generator instead: it is the same speech pass written out as SRT or VTT.
Which languages can it transcribe, and how do I get a better result?
Eighteen, and Auto detect if you would rather not choose: English, Spanish, French, German, Italian, Portuguese, Dutch, Polish, Russian, Ukrainian, Turkish, Arabic, Hindi, Urdu, Chinese, Japanese, Korean and Vietnamese. Naming the language beats leaving it on Auto detect when you already know it, because detection is one more thing that can go wrong. Then use the Names or special terms box, which holds up to 400 characters, for the people, brands and technical words the recording uses, so the model has them before it starts rather than guessing at them. Clear speech with little background noise produces the best result.
What happens to my video after transcription?
Your video, extracted audio, and transcript stay on TryDeputize infrastructure and are deleted automatically within about an hour. A video waiting in the transcription queue is kept only until it runs, with a 12-hour safety limit for abandoned work. Speech is processed by our self-hosted faster-whisper worker, so your recording is not sent to a third-party AI service. That worker has a queue of its own, kept apart from the one the PDF and image tools use, so a long transcription never slows those down and a rush on them never delays yours.