Add captions to Facebook videos
Drop a clip above and get clean, animated captions tuned for Facebook. An on-device AI model transcribes your speech with word-level timing, every word stays editable, and you export a 1:1 MP4 with the subtitles burned in — or an SRT file — with no watermark, no account, and nothing ever uploaded.
Caption a Facebook video the way the feed actually plays it
Facebook posts sit in a feed that keeps scrolling, autoplaying one clip after another. A viewer decides whether to stop in the first second or two, often before any sound kicks in. Captions turn that silent autoplay into something readable, so your point lands while the clip is still on screen rather than after a tap that may never come.
The tool above handles this end to end. Load your footage, let the AI write the transcript, pick a caption style, and place the words so they read cleanly inside a 1:1 frame. There is no Facebook login involved and nothing to schedule — you are just making the file, which you can then upload to a Page, a profile, a group, or Reels.
Why sound-off captions matter on Facebook
Most Facebook video is watched with the sound muted. That is the default for nearly everyone who sees your post, not an edge case to plan around. If the message lives only in the audio, most of your audience never receives it, however good the clip is.
Captions close that gap. They carry the narration, the names, the numbers, and the punchline for the silent majority, and they make the video work for anyone who is deaf or hard of hearing. On a platform with this much muted viewing, captions belong in the edit rather than tacked on after.
Word-perfect timing from an AI that reads ~99 languages
The speech model listens to your clip and transcribes it with word-level timestamps, so each word appears exactly when it is spoken instead of drifting a beat behind. It auto-detects the language across roughly 99 of them, and you can override that choice or regenerate the whole transcript in another language for a different Facebook audience.
Nothing here is locked. Every word in the transcript is editable, so you can fix a name, tidy a filler word, or correct a brand spelling before export. If a line lands a fraction early or late, nudge its timing until it sits right against the audio.
Private by construction: the file never leaves your device
This runs entirely in your browser. Your video is read locally, the transcription happens locally, and the export is rendered locally. The file is never sent to a server, never uploaded, and never used to train any model. That is a guarantee built into how the tool works, not a line in a policy.
The AI model downloads to your browser once, then runs offline on your machine for every clip after that. For a behind-the-scenes cut, a customer testimonial, or anything you would rather not hand to a third party before it is even posted, the footage stays with you the whole time.
Caption styles that suit a square, broad-audience feed
There are four animated styles. Karaoke highlights each word as it is spoken, Highlighted drops the active word into a colored box, Minimal shows one clean word at a time, and Dynamic gives that single word a small pop. In a 1:1 frame the square crop lets captions sit a touch higher than a vertical clip would, so they stay clear of the feed's overlaid buttons and the post text below.
You control the rest: typeface from Inter, Montserrat, Oswald, Lora, or JetBrains Mono, plus weight, size, top, center, or bottom position with fine nudging, text and highlight colors, outline, shadow, and words per line. Facebook's audience is wide and mixed, so a slightly larger size with a solid outline keeps the text legible on small phones and crowded feeds alike.
Export a 1:1 MP4 with captions burned in
When the timing looks right, export an MP4 with the captions burned straight into the picture at 1:1, 1080 by 1080 — the square format Facebook serves cleanly in the feed. Burning the words into the frame means they travel with the video everywhere it gets shared and reposted, which matters on a platform built around sharing. Rendering is hardware-accelerated, and you can optimize for sharing size or hold source quality. There is no watermark.
Prefer separate subtitle files? Download an SRT or VTT and upload it alongside your video so Facebook shows toggleable captions. Facebook handles long content too — up to 240 minutes — so for a full talk or stream, the editable transcript and per-line timing keep even a lengthy caption track accurate from start to finish. Every style, language, and export option is free for everyone, with no sign-up and no paywall.
Questions
Not always, but burned-in captions are the safer choice. They show no matter how the clip is viewed or reshared, and they survive being downloaded and reposted, which happens a lot on Facebook. Export a 1:1 MP4 with the words rendered into the frame. If you would rather keep captions toggleable, download an SRT or VTT here and upload it with your video instead.
Aim for text that is easy to read on a phone in a moving feed: a generous size with a clear outline or shadow so it holds up against any background. In the square 1:1 frame, keep captions out of the very bottom, where the feed stacks the post caption and action buttons. The position and nudge controls let you set them at a comfortable height, and you can preview before exporting.
Yes. Every caption style, all ~99 languages, and both export paths — burned-in MP4 and SRT or VTT — are free for everyone, with no sign-up, no account, and no watermark on the output. Free here is a deliberate choice about how the tool should work, not a limited trial.
No. Everything runs in your browser on your own device. The video is never uploaded, never sent to a server, and never used to train any model. The AI model downloads once, then transcribes locally for every clip after that, so your footage stays with you the entire time.
Yes. Facebook allows clips up to 240 minutes, and the tool transcribes the whole thing with word-level timing. Because the transcript is fully editable and you can fine-tune any line's timing, even a long caption track stays accurate — fix a word, adjust a line, and export when it reads right.