🎬 Auto Captions for Video

How to add auto captions to a video

Step 1: Upload your MP4, WebM or MOV file above — it never leaves your device; everything runs in your browser.
Step 2: Choose the spoken language and press Generate Captions. The AI listens and writes timed subtitles.
Step 3: Pick one of 7 styles — from classic subtitles to viral Hormozi and TikTok pop-up captions — then download.

Why add captions at all?

Most social video is watched on phones, and a large share of it starts muted — in feeds, on trains, in quiet offices. Captions keep those viewers watching instead of scrolling past, and they make your content accessible to deaf and hard-of-hearing audiences too. Creators also find that captioned clips hold attention for longer, because viewers can follow the words even when they cannot hear the voice.

This tool is built for exactly that job. The speech recognition runs on your own device, so a talking-head clip, a product demo, a lecture recording or a family video can be captioned without uploading it anywhere. Choose a quiet, classic subtitle style for formal content, or one of the bold word-by-word styles that short-form creators use to stop the scroll. If a name or a technical term comes out wrong, fix the line in the editor before you export — then download the finished video with the captions burned in, or take the .srt file to use on YouTube and other platforms.

The 7 caption styles, explained

Every style renders the same timed words — only the look changes:

• Classic — clean white subtitles with a dark outline. The safe choice for lectures, interviews and anything formal.
• Hormozi — big, bold capitals with the spoken word highlighted in yellow. Built for short-form hooks on TikTok, Reels and Shorts.
• CapCut Box — bold capitals on a solid black box, with the spoken word highlighted in an orange box, word by word. The signature CapCut look used by story and explainer channels.
• Karaoke — the line stays on screen while the current word lights up, so viewers can follow along word by word.
• Pop-up — one word at a time, centre screen, with a punch-in effect. Maximum emphasis for punchy, fast-cut clips.
• Minimal — small, light text on a soft dark band. Understated and professional for business or brand videos.
• Stroke — heavy Impact-style lettering with a thick outline. The meme and commentary look.

Who uses this tool?

Short-form creators captioning TikToks, Reels and Shorts so muted scrollers stop and watch. Teachers and course creators making recorded lessons accessible. Small businesses adding captions to product demos and ads without paying per minute. Anyone with a family video, a sermon, a podcast clip or a meeting recording who wants the words on screen — without handing the file to a stranger's server.

Tips for the most accurate captions

• Start with clear audio. Captions are only as good as what the AI can hear — quiet room, close microphone, one speaker at a time.
• Pick the exact language spoken in the video, not the language of the audience.
• Choose the right model. "Fast" is fine for clear English; "Accurate" is the better default; for Urdu and Arabic, the largest model ("Best for Urdu/Arabic") is worth the bigger one-time download.
• Fix names and jargon in the editor. Proper nouns and technical terms are what AI hears worst — a two-minute pass in step 4 makes captions look hand-made.
• Size for phones. Most viewers watch on a small screen: Large font, bottom position, and a bold style survive compression and small displays best.

Honest limitations

The speech-recognition model downloads to your browser the first time you use the tool — about 150 MB for the standard models, up to roughly 1 GB for the largest — and then stays cached, so later videos work offline. Because everything runs on your own device, a long video on an older laptop takes a while, and the export step renders in real time (a 10-minute video takes about 10 minutes to export, saved as a WebM file). Accuracy drops with loud background music, overlapping speakers and strong accents; that is exactly why every line stays editable before you download.

Frequently asked questions

Is it really free?

Yes. No signup, no watermark, no limits. The AI model downloads once to your browser (about 150 MB), then works offline.

Which languages are supported?

English, Urdu, Hindi, Arabic, Spanish, French, German, Portuguese, Turkish, Persian and more. Accuracy is highest in English; for other languages we show an honesty notice before you start.

Is my video uploaded anywhere?

No. Transcription and caption rendering happen entirely on your device. Nothing is sent to any server.

Can I edit the captions?

Yes — every line is editable before you export, and you can also download a standard .srt subtitle file.

Which caption style should I choose?

For TikTok, Reels and Shorts, Hormozi, CapCut Box, Karaoke and Pop-up hold attention best. For lectures, interviews and business videos, Classic or Minimal look right. You can switch styles after transcribing and preview each one before exporting — the words do not change.

How accurate are the captions?

For clear English speech, around 90% of words come out right. Background music, crosstalk, strong accents and quiet recordings lower that. Urdu and Arabic do noticeably better on the largest model. Either way, review the lines in the editor — fixing the handful of mistakes takes a couple of minutes.

How long does it take?

The first use includes the one-time model download. After that, transcription runs on your device, so speed depends on your hardware — short clips usually finish quickly on a modern laptop, while long videos take longer. Exporting renders the finished video in real time.

What exactly do I download?

Two things, your choice: the video with captions burned in (WebM), ready to post anywhere, and a standard .srt subtitle file you can upload to YouTube or any platform that accepts subtitle files.