Turn a video file into SRT subtitles in its own language
Upload a video or audio file, or paste a Google Drive or Dropbox link, and get timed subtitles in the language spoken, as SRT and VTT. A video of up to 15 minutes is free in full. A longer one shows its first part free on screen, and the whole video is $5, paid once, for up to 6 hours.
- Subtitles in the language spoken: nothing is translated in this mode
- Cues of at most two lines of about 42 characters, on screen for 7 seconds or less
- SRT and VTT downloads in UTF-8, ready for YouTube Studio, VLC, Premiere and DaVinci Resolve
- One file up to 2 GB; a public Google Drive, Dropbox or direct file link also works
- Up to 6 hours for $5; Unlimited covers up to 12 hours per video
How to make an SRT from a video file
-
1
Add your video
Drop the file in the box, up to 2 GB, or paste a public Google Drive, Dropbox or direct link. YouTube links are not taken here.
-
2
Keep "Only the language spoken"
This page preselects it. Choosing the spoken language is optional, but it helps when the video opens with music or applause.
-
3
Check, pay once, download
Read the free part on screen, pay $5 for the whole video if it looks right, correct any word on screen, then download SRT or VTT.
Why a long video is hard to subtitle in a chat assistant
A chat assistant can often caption a two-minute clip well. A two-hour conference session is another matter. Assistants that accept audio limit how long a file can be, free plans often cannot upload audio at all, and a full set of subtitles for a long video does not fit in one reply. The usual result is a file that stops partway, or a summary where the subtitles should be. Timing is the other weak point: an assistant cannot reliably say when each sentence starts in a long file, and an SRT whose times drift is worse than none.
Here the video is cut into parts of up to ten minutes. Each part is listened to in full and written as timed subtitle cues, the times are shifted by where that part starts, and the parts are joined in order. Code then checks that the cues run in order and do not overlap, instead of trusting the model to get that right.
What one cue looks like
- A sentence-sized piece of speech, so the viewer is not left mid-thought
- At most two lines of about 42 characters each, a common reading length for subtitles
- On screen for no more than 7 seconds
- Words in the language and script spoken; a speaker who switches to English mid-talk is written in English
- Unclear words marked [inaudible] instead of guessed
Recordings that fit this page
- Conference sessions, keynotes and panels of one to three hours
- University lectures and seminars posted for students
- Webinars and training videos for an internal library
- Interviews and documentaries where you want same-language captions for accessibility
Honest limits
The timings come from the AI listening to the audio. They are close to the speech but not frame-exact, so if you deliver broadcast captions, open the SRT in a subtitle editor and nudge the cues that matter. Crowd noise, music under speech and people talking over each other cause mistakes, so read names and numbers before you publish. The file is a working draft made by an AI speech model, not a certified caption file.
How the tool works, what it costs and how your files are handled is on the Video to Translated Subtitles page; prices for every tool are on the pricing page.
Last updated
Questions and answers
Can ChatGPT make an SRT file from my video?
For a short clip it often can. For a video of an hour or more, the upload limit, the reply length and unreliable timestamps tend to get in the way, and the result usually stops partway or loses its timings.
How much does a 2-hour lecture cost?
$5 for the whole video, paid once, for anything up to 6 hours. A video of up to 15 minutes is free in full. Unlimited, $19 a week, covers up to 12 hours per video; cancel any time.
Where can I use the SRT?
Upload it with timing in YouTube Studio under Subtitles, load it in VLC with Subtitle > Add subtitle file, or import it as a caption track in Premiere or DaVinci Resolve. Plex reads it too.
Can I get a translation as well?
Yes. Change the output to Translated or Bilingual and pick a subtitle language. A translated job also includes an SRT in the language spoken, so you end up with both.
Who can see my video?
Only you. It is sent to an AI model provider to be processed and is not used to train models. Jobs started without an account are deleted after 7 days, and you can delete a job yourself at any time.