Subtitle generator
Load a video and let the speech recognition write the subtitles for you — then fix any line by hand, download the result as an .srt or .vtt file, or burn the text straight into the picture. Everything runs on your own device; your video never leaves your computer.
How this subtitle generator works
A speech-recognition model runs inside your browser and listens to your video's audio track — nothing is uploaded, so an interview, a lecture or a family film stays on your machine. It writes a first draft with start and end times already in place; you read it over, fix a name or a technical term where needed, and choose how the subtitles should leave: as a separate .srt or .vtt file that platforms like YouTube accept directly, or burned into the picture so the text plays everywhere, including apps that ignore subtitle files. The preview player above the editor shows your lines exactly as a real player renders them, updating live while you type.
How good are the automatic subtitles?
Good enough to be a real time-saver, honest enough to need a read-through. Clear speech comes out largely word-perfect with usable timings; names, jargon and people talking through each other come out wrong more often. Treat the result as a solid first draft: skim every line before you publish — that is exactly what the editor under the player is for.
Which languages does it understand?
The speech model understands dozens of spoken languages — the list covers the common European ones and detects the language automatically if you leave the selector alone. There is also a translate option that turns any spoken language directly into English subtitles in one pass. The subtitle text itself you can of course edit into any language you like.
SRT or VTT — which one do I need?
SRT is the older, universal one: video players, editing software and most upload forms accept it. VTT is the web-native format that HTML5 players and browsers use. The content is the same subtitles either way, so when in doubt take SRT — and if a website specifically asks for WebVTT, take VTT. Both downloads come from the very same lines you edited.
Should I burn the subtitles in or keep them as a file?
A separate file is the flexible choice: viewers can switch subtitles on and off, platforms index the text, and you can correct a typo without re-encoding the video. Burning them in makes the text part of the picture — it plays absolutely everywhere, which is why social media clips are usually burned. Many creators do both: upload the file where possible, burn for the platforms that ignore files.
Is my video uploaded anywhere?
No. The speech recognition, the editing and the burning all happen inside your browser on your own device. The first time you press the generate button, the speech model downloads once (about 77 MB) and your browser keeps it after that — that download is the model coming to your video, never your video going anywhere.
Can I make subtitles in several languages for one video?
That workflow lives in the CBCOM Video Studio: there one set of timings carries any number of languages, so you transcribe once and type the translations against the same times. This page deliberately keeps to one subtitle track — one video, one language, done.
Adding subtitles to a video, step by step
Subtitling used to mean either paying a service by the minute or typing timestamps into a text file by hand. Speech recognition changed that: today the honest workflow is generate, correct, export — let the machine do the tedious ninety percent and spend your own attention on the ten percent it gets wrong. That is precisely how this page is laid out, from top to bottom.
Why subtitles are worth the effort
The large majority of short videos on social media are watched with the sound off — without text, your message simply does not arrive. Subtitles also open your video to deaf and hard-of-hearing viewers, to anyone watching in a second language, and to search engines, which can only read text. A subtitled video is watched longer, understood better and found more often; few minutes of work pay back on every single view.
Getting a clean first draft
Pick the spoken language yourself instead of relying on automatic detection when you know it — it removes one source of error. Speech over loud music, several people talking at once and heavy dialect are the hard cases; expect to correct more there. If your video is in another language and you want English subtitles, tick the translate option and the model does both jobs in one pass.
Editing like a professional
Keep lines short — one to two lines on screen, roughly forty characters per line — and let no line linger less than a second or the eye cannot catch it. Read the text against the player above the editor: it renders your lines exactly as a real player will. Fix names and technical terms first (they are the most common mistakes), then check that sentences do not break in awkward places. The times accept 1:23.45 as well as plain seconds, and a comma works as the decimal sign.
Choosing the right export
Uploading to YouTube or another platform with subtitle support? Download the .srt and upload it next to your video — viewers keep the choice to switch it on or off, and you can replace the file any time. Posting to a platform that ignores subtitle files, or sending the clip around directly? Use the burn-in button: the tool plays your video through once in real time and records the text straight into the picture, so it shows everywhere, on every screen, forever.
