Ember

Free · Unlimited · Runs in your browser

Captions that pop.
No minutes to buy.

Drop in a video. Ember writes down every word, times each one to your voice, animates them in the style you pick and burns them into an MP4. It all happens on your own computer, so there’s no upload, no queue and no meter running.

  • ✓ No sign-up
  • ✓ No watermark
  • ✓ No length limit
  • ✓ Nothing uploaded
Made for TikTokInstagram ReelsYouTube ShortsYouTubeLinkedInPodcast clips

The per-minute problem

Caption apps run a meter.
Ember doesn’t have one.

Online caption tools upload your video and process it on their servers, and servers cost money by the minute. So they sell you minutes: a monthly allowance, a cap, and an upgrade screen the week you post more than you planned.

Ember brings the tools to your video instead. Your own computer does the listening, the drawing and the encoding, so captioning a video costs nothing and there’s nothing to count.

A metered caption plan Resets in 9 days
113 of 120 minutes used Upgrade to keep posting
Ember Never resets, never runs out
∞ minutes, with no account
and no watermark

A 15-second clip or a 15-minute interview. Same price: nothing.

How it works

From raw clip to captioned MP4 in about a minute.

Try it now →
  1. 01

    Drop in a video

    MP4, MOV or WebM, straight from your phone or your editor. Vertical, square or wide.

  2. 02

    Ember writes every word

    Whisper, the open speech model, runs on your graphics card and writes down each word with the moment you said it.

  3. 03

    Style it and export

    Pick a look, fix any word with a click, drag the captions where you want them, then export the MP4. Subtitle files too.

Styles

6 looks to start from. Every detail yours.

Change the font, colours, size, outline, animation and how many words show at once. What you see in the editor is exactly what gets burned into the video, frame for frame.

Private by design

Your video never leaves your computer.

Ember is a web page that does the work itself. The speech model downloads once, from about 80 MB, and your browser keeps it. Listening, drawing and encoding all happen on your machine. Nothing is sent to us or to anyone else.

  • ✓Unreleased videos stay unreleased.
  • ✓Client work stays on your machine, under your control.
  • ✓No queue behind other people’s videos. Your computer is yours.

The details

Everything you’d pay for, minus the bill.

Word-by-word timing

Each word lights up the moment it’s said, not a sentence at a time.

Fix anything in a click

Click a caption and retype it. Names, slang, brand words. The timing follows your edit.

Put captions anywhere

Drag them above a face or below a product. Ember keeps them on screen.

Ums and uhs, gone

Strip filler words from your captions in one click.

Dozens of languages

Spanish, Hindi, Portuguese, Japanese and many more, detected automatically.

Subtitle files too

Download SRT or VTT for YouTube and your editing app, alongside the video.

Questions

Before you drop one in.

Something else? support@anvilhq.dev

Is it really free?

Yes. Your computer does the work, so a video costs us nothing to caption and we don’t charge for it. No watermark, no account, no limit on how long or how many. Paid extras for heavy users, like batch export and saved brand styles, may come later.

Does my video get uploaded anywhere?

No. The video, its sound and its words stay in your browser. The page downloads the speech model from Hugging Face the first time, and when you export, it tells us how long the video was and which style you used, so we can see Ember is being used. That’s all we receive.

What do I need?

A laptop or desktop with a recent Chrome or Edge works best, because Ember can use your graphics card. Recent Safari and Firefox work too. Without a usable graphics card, transcription runs on the processor and takes longer.

How long does it take?

On a laptop with a graphics card, a minute of speech usually takes well under a minute to transcribe, plus a one-time model download the first time. Exporting is usually faster than watching the video.

How accurate is it?

Ember uses OpenAI’s Whisper, one of the most accurate open speech models. Clear speech comes out nearly perfect with the Fast model; for accents, background noise or other languages, choose Accurate or Best. Either way, every caption is editable before you export.

Which files does it take and make?

MP4, MOV and WebM in. Out comes an MP4 (H.264) at your video’s own resolution with its original sound, plus SRT, VTT and plain-text transcripts if you want them.

Can I use it for client work?

Yes. What you export is yours to use however you like.

Your next video, captioned in about a minute.

No account, no upload, no meter. Just drop it in.

Caption a video, free