Free auto captions: animated word-by-word captions, burned into your whole video, in any of 37 languages, with no sign-up. Or let it cut the strongest 30 or 60 seconds, framed on you and ready to post. Your video never leaves your device. Free to use: two videos every 30 days with no mark, or $49 once for unlimited videos.
Free for two videos every 30 days, no sign-up, no watermark.
Carry on where you left off
Live preview, rendered on this page by the same engine that burns your video.
Your video stays on this phone either way. How should the words be written?
Asked once; change it any time under Listening settings.
Which language is spoken in this video?
Your own or a client's logo on the video, and saved brand presets, come with Pro.
Brand presets
A preset keeps the caption style, its colours, position, size, words at a time and the logo. Presets and logos stay in this browser and are never sent to us.
The same video again, with the other captions. It does not count as another video.
Subtitle files are free for any recording up to ten minutes, and follow the cut when you choose the best 30 or 60 seconds.
Words this browser remembers to fix
Keys: Space plays and pauses, the arrow keys move 5 seconds (1 with Shift).
Waiting for the transcript.
Heard as The standard speech model often writes Hindi in Urdu letters.
Every word as it was heard. Click a word to correct it and the captions follow. Click the time to play from there.
Burncaps is a browser tool that burns animated word-by-word captions into your video, or cuts a talking video to its strongest 30 or 60 seconds. Your video opens in the page, the AI listens on your device, and you choose what comes out.
Your file opens directly in the page and the video is never uploaded, which you can verify in your browser's network tab. OpenAI's open-source Whisper model runs in your browser and times every single word. On a phone, for a language other than English, or when you ask for a closer transcript, you can choose our speech service instead: faster and more accurate, and it gets only the sound, never the picture.
The whole video: animated word-by-word captions burned in, nothing cut. Keep your room as it was, or blur it, put a studio colour or your own picture behind you, or a green screen to key out later. The best 30 or 60 seconds: ums and dead air cut, the moment chosen from your transcript and ended where your voice actually finishes the point, your product names spelled right, a 9:16 window placed on your face, the background softened or swapped for a studio black, white, grey or a colour of your own, an opening card, and a beat at the end that fades instead of stopping.
Your transcript opens as a script under the video. Click any word to correct it (a product name the model has never heard, say), fix it everywhere in one click, and the captions follow. Pick one of twenty caption styles and your browser encodes the finished file, usually under a minute. Subtitle files come free with every video up to ten minutes; a recording up to 90 minutes gives its script and subtitle files on a computer, the first ten minutes free and all of it with Pro.
All of this changes the script and the captions without listening again, on every free video as well as with Pro. Correct any word by clicking it, and one click changes every other place the same word was heard. Add words the listener missed. Cut a word by emptying it (in the best 30 or 60 seconds its sound goes too). Show one to six words at a time. Move the captions up or down and make them bigger or smaller. Set the colour of the text and of the word being said, over any of the twenty styles, or let each word appear as it is said. Type the names to spell right, and this browser remembers them, and every fix you made everywhere at once, for your next video. Phones export too: in the seven days to 8 October 2026, phones finished at least 29 exports and none failed (our own tests not counted; checked 8 October 2026).
No B-roll, no stock music, no picking twenty clips from a two-hour podcast. It is for one recording of up to ten minutes: caption it, or cut it to its best half minute. A recording up to 90 minutes, on a computer, gives its script and subtitle files instead. Export is fastest in Chrome or Edge on a computer; many phones export too, more slowly.
Word-by-word highlight animation, the format proven to hold attention on short-form video, in the looks creators use most right now: keyword highlights, the native text box, words that ink in, tall capitals, a sliding underline and more. Every style is included in every video, free or Pro. Every thumbnail below is rendered live by the engine.
Free to use: two videos every 30 days with no mark, so you can judge the result on your own footage; one payment gives unlimited videos, for good. Pro is never a subscription.
Get Pro for $49, once“It is an exceptionally good product and I’m happy with it so far.”
Zaden, United States, a Pro buyer
| What matters | Burncaps | Typical caption app |
|---|---|---|
| Price | Free to use: two videos every 30 days with no mark, or $49 once for unlimited videos | $10 to $25 a month |
| Watermark on free videos | None on the two free videos every 30 days, and none with Pro | Usually, from the first video |
| Your video on their servers | No. It stays on your device | Yes |
| Account required | No | Usually |
| Free videos | Two every 30 days with every feature and no mark | Minutes-per-month caps and a watermark |
Nothing for two videos every 30 days, with every feature included: both outputs, all twenty styles, every background, the max-accuracy model, no account, and no watermark. They are counted per browser and per internet connection, so on a shared network (an office or a university) someone else may already have made this month's two; signing in with your email then gives you your own two. Pro, $49 once, gives unlimited videos with no mark, and Pro is not a subscription. Subtitle files are free without limit for videos up to ten minutes, and for the first ten minutes of a longer one.
The whole video keeps every second you recorded and burns animated captions in; you can leave your room as it was, blur it, or put a studio colour, your own picture or a green screen behind you. The best 30 or 60 seconds is an edit: ums, stutters and dead air are cut, the strongest stretch is chosen from your transcript and ended where your voice finishes the point, your product names are spelled right, a 9:16 window is placed on your face, the background is softened or replaced with a studio backdrop, an opening card carries your line, and the ending holds for a beat and fades instead of stopping. You can switch between them after the transcript arrives.
Yes. Choose a shape for the whole video: Original, Vertical 9:16 for TikTok, Reels and Shorts, Square 1:1, or Portrait 4:5 for the Instagram feed. Follow me keeps you in frame like a calm camera operator: it stays still while you stay near the middle and moves smoothly when you step away. Fixed lets you drag the picture into place and zoom in. Fit everything shows your whole frame with nothing cut, for slides, screen recordings or two people. Every caption style and background works in every shape, and a landscape take cut to vertical is enlarged, so it is a little softer than the original.
Yes. Your transcript opens as a script under the video. Click the word, type the right one, and the captions and subtitle files follow. If the same word is wrong elsewhere, one click changes them all, and this browser remembers the fix for your next video. Typing your product and company names into the names box before you start catches most of them up front.
Everything below changes the script and the captions without listening again, on every free video as well as with Pro. Correct any word, and one click changes every other place it was heard. Add words the listener missed. Cut a word by emptying it (in the best 30 or 60 seconds its sound goes too). Show one to six words at a time. Move the captions up or down and make them bigger or smaller. Set the colour of the text and of the word being said, over any of the twenty styles, or let each word appear as it is said. Names to spell right are remembered by this browser for your next video.
No. The model downloads into your browser; your video never leaves it. The sound can leave only if you choose our speech service, which a phone offers for speed and a computer offers for a language other than English, for a slow computer, or when you ask for a closer transcript: then it goes in short pieces to Microsoft's MAI-Transcribe in the EU, never the picture, and nothing is kept. Before and after an export the page makes one small check with our server, carrying a random ID for your browser and nothing from the video (only a finished export uses a free video), and each page asks once which price applies where you are, carrying nothing about you. The transcript text can leave only when you choose the 30 or 60 second cut, type names to spell right, or choose a caption language, sent to our AI to choose the stretch, fix the names or translate the lines; it is not stored, and ticking "Keep everything on this device" stops it. An email you type, to hear from us (we first send one email to confirm it is yours) or to send yourself this page, goes to our server and nothing else from the page goes with it. You can verify all of this in your browser's network tab.
A recent browser. Chrome or Edge on a computer is the fastest and surest, because they use the graphics chip to write the video; Firefox may save a WebM file instead of an MP4, and many phones export too, more slowly: in the seven days to 8 October 2026, phones finished at least 29 exports and none failed. Where a browser cannot write video, you can still transcribe and download SRT, VTT or ASS files.
37, the list in Listening settings. It listens first and names the language it hears above your script, with a one-click change. Twelve Indian languages, Hindi, Urdu, Bengali, Tamil, Telugu, Gujarati, Marathi, Kannada, Malayalam, Punjabi, Nepali and Odia, can be written by our speech service (Microsoft's MAI-Transcribe, which names the language by itself and gets only the sound) or on your own device by a speech model made for each (AI4Bharat IndicConformer, open source, 138 MB once per language); Burncaps asks once which you want. On Google FLEURS test recordings the service got about 90 in 100 words right in Hindi, 85 to 87 in Urdu, Bengali, Gujarati and Punjabi, and 79 to 83 in Tamil, Telugu, Marathi, Kannada, Malayalam, Nepali and Odia (the full test); the on-device models got about 89 in Hindi, 84 in Bengali, 72 to 81 in most others, 68 in Nepali and 60 in Malayalam. Accuracy is best in English and the major European languages; for a strong accent, Max accuracy in Listening settings got about three times fewer words wrong in our test on Indian-accented English. Sinhala and Assamese speech is not written well by any speech model that runs in a browser yet, and Burncaps says so rather than guess; captions in them on an English video work, through Captions in. Pages for Hindi · Urdu · Bengali · Tamil · Telugu · Gujarati · Marathi · Kannada · Malayalam · Punjabi · Nepali · Odia · Spanish · Portuguese · Arabic · Indonesian · Tagalog · French · German.
The speech model downloads once (about 190 MB on a computer with a graphics chip, about 80 MB without one) and is kept by your browser. Burncaps starts fetching it the moment you reach for a file, and after the first time the words start appearing on your video within seconds. A video in one of the twelve Indian languages also downloads the model made for that language, 138 MB once, unless you choose our speech service, which needs no download at all; on a phone the service answers in seconds, where a phone’s own model takes minutes. The max-accuracy model is a 1.4 GB download, once.
Unlimited videos and exports with no mark, and no daily, monthly or lifetime cap. Video with captions burned in: up to ten minutes and 1.2 GB per file on a computer, the same as free. A longer recording, up to 90 minutes on a computer: the script and subtitle files of all of it, where free gives the first ten minutes. Translations and best-moment picks are unlimited for your own use, with safety limits per account, for a day and for any 30 days, that we do not publish; the daily one resets at midnight UK time. Five hours of our speech service in any 30 days, then your own device listens. Three devices at once. Features added later come with it, and Pro is not a subscription. Nothing in Pro today is ever moved into another plan. Every limit is set out in plain words in About.
Burncaps is built by Jordan Labs Ltd, a UK company (no. 17358646). No ads, no cookies, no tracking of individuals. Measurement is a cookieless visit counter and anonymous step counts with no identifier, listed in the privacy policy: each carries where your visit and your first visit came from, and the one made when you drop a video on one of our other pages also says which page. Paying customers are the entire business model: Pro, one payment, and Studio, a monthly plan for teams and agencies.
What it is: a browser tool that burns animated word-by-word captions into a talking video, or cuts it to its best 30 or 60 seconds.
Price: free with every feature, for two videos every 30 days with no mark; Pro, $49 once, gives unlimited videos with no mark. Pro is not a subscription.
Privacy: the video never leaves your browser. The sound leaves only if you choose our speech service.
Files: MP4 or MOV up to ten minutes in; an MP4 with the captions burned in out. SRT, VTT, ASS and text files are free without limit for recordings up to ten minutes. A longer recording, up to 90 minutes on a computer, gives its script and those files: the first ten minutes free, all of it with Pro.
Languages and styles: captions in the language spoken, twelve Indian languages heard on your device, translation into 37 languages, 20 caption styles.
After it listens: correct any word (one click changes it everywhere), add or cut words, one to six words at a time, position, size, text and highlight colours, and names this browser remembers, all free and without listening again.
Made by: Jordan Labs Ltd, UK company no. 17358646.
Against other tools: sixteen caption tools compared, from the vendors' own pages.
Pro is $49, once, for unlimited videos: captions on the whole video or the best 30 or 60 seconds, every style, the max-accuracy model, and every future feature. Pro is not a subscription.
Payment by card through Stripe. 14-day refund if it does not work on your machine. Subtitle files stay free either way.
Not today? One tap tells us why.
Thank you. It helps us set the price right.
Use the email you paid with. We send a six-digit code to it. No password.
Not bought Pro yet? Burncaps is free to use. Get Pro for $49, once.
Signed in as on this device. Unlimited videos.
Pro works on up to three devices at once. Signing in on a fourth signs out the one used longest ago. On another device, choose Pro sign-in and use this email.
Studio works on up to ten devices at once. Signing in on an eleventh signs out the one used longest ago. On another device, choose Pro sign-in and use this email.
Manage or cancel StudioTo cancel Studio, email hello@burncaps.app from this address. It stays on to the end of the month you paid for.
Signed in as .
Get Pro for $49, once