How to make auto captions in After Effects without uploading your audio
After Effects has no speech-to-text of its own; Adobe's transcription lives in Premiere Pro. To make auto captions inside After Effects you need a panel or plugin. Caption studio in the mouzeclick workspace transcribes the layer on your own computer, in 100 languages plus Auto Detect, lets you fix every word, then builds styled, animated caption layers. It runs on Windows.
- Transcribed on your computer: the Caption studio dialog says no data leaves your machine
- 100 languages (Hinglish and 99 named ones) plus Auto Detect, with offline translation to English on the larger models
- Eight speech engines, from about 420 MB to about 4 GB, each downloaded once; Whisper Large V3 Turbo is the default
- 1 to 15 words per line or no limit; six ready looks, five layouts, 13 word animations and keyword highlighting
- Out as styled caption layers, subtitle layers from an SRT written for you, or a JSON export. Windows only
Can After Effects make auto captions on its own?
No. Adobe documents Speech to Text as a Premiere Pro feature: transcription in 18 languages, translation of captions into 27, and offline use once a language pack is installed. After Effects has no transcribe command.
So there are two routes. The Adobe-only route is to transcribe in Premiere Pro, fix the transcript there, export the captions as an SRT file, then get that text into After Effects, which takes a script or a lot of copy and paste. It works, and if the video is cut in Premiere anyway, captioning there is the sensible call (see auto captions in Premiere Pro).
The other route is a panel that transcribes inside After Effects, so the captions are born as layers in the comp you're animating. That matters for motion work such as kinetic captions, a word that punches on the beat, or a keyword that changes colour. Doing that from an imported SRT means rebuilding every line.
How do you add auto captions in After Effects with Caption studio?
Transcribe, read it through, then build. Caption studio is never a press-and-forget button, and it shouldn't be.
- Open Window › Extensions › mouzeclick workspace. Caption studio opens as a dialog; the Overview page also has an Auto captions shortcut.
- Choose the source: the selected layer, every media layer in the comp, or a file.
- Choose the language. Auto Detect is the default.
- Choose the speech model (next section).
- Choose the range: the full layer, the work area only, or the layer's span only.
- Set words per line, from one to fifteen, or no limit.
- Transcribe. Recent transcriptions are kept, so you can come back to one.
- Read it through. Every word is a chip: double-click to fix it, click to make it a keyword, with undo and redo.
- Style it in the Style, Text and Motion tabs, with a live preview.
- Build: styled caption layers, standard subtitle layers from an SRT written for you, or a JSON export.
No audio handy? Load Demo Transcript tries the styles on sample text. And if the clip has long pauses, run the Silence remover in Motion lab first (Dialogue, Podcast or Aggressive), so you caption the tight cut and never retime a line.
Which languages do After Effects auto captions support in Caption studio?
Hinglish and 99 named languages, plus Auto Detect, so 101 choices in the menu. Hindi can be written in Roman letters, the way people type it on their phones, and the larger models translate captions to English offline.
Counts differ between mouzeclick's caption tools, so check which one you're reading about: mouzeclick Studio in Premiere Pro has 47 languages, the Premiere Pro workspace's Caption studio has 40 options, and the After Effects Caption studio has 100 plus Auto Detect. For Hinglish on a Premiere timeline, see Hinglish captions in Premiere Pro.
Auto Detect is good on clean, single-language speech. Pick the language yourself for clips under ten seconds, for strong accents and for anyone who switches languages mid-sentence; that's where detection guesses wrong.
Which speech model should you choose for captions?
Start with Whisper Large V3 Turbo, the default; it suits most talking-head work. Caption studio offers eight speech engines, each a one-time download that then runs on your computer:
| Engine | Download | Use it for |
|---|---|---|
| Whisper Large V3 Turbo (default) | about 1.6 GB | Everyday YouTube, podcasts and reels |
| Whisper Large V3 | about 3.0 GB | Noisy rooms, heavy accents, work going to a broadcaster |
| Whisper Medium | not stated | Older laptops, rough transcripts you'll rewrite anyway |
| Moonshine Medium | about 420 MB | The smallest download |
| Qwen3-ASR 0.6B, NVIDIA Canary 1B v2 | about 3.8 GB, about 4 GB | A second opinion when Whisper struggles with a clip |
| IndicConformer, SraVaani 1.0 | about 2.4 GB, about 900 MB | Engines built for Indian languages |
With an NVIDIA graphics card, Whisper can use the GPU to run faster. Download only the engines you'll use, because together they come to more than 15 GB.
How many words should a caption line have?
As few as the format allows. Caption studio holds 1 to 15 words per line, or no limit. What has worked for me:
| Format | Words per line | Why |
|---|---|---|
| 9:16 shorts and reels | 1 to 3 | Big type, read at a glance, timed to the beat of speech |
| 16:9 talking head for social | 4 to 6 | One line at the bottom, no second line covering the chin |
| Long-form subtitles | 8 to 12 | Reads like a sentence; fewer flashes on screen |
For vertical video, check where the platform's buttons land before you place the captions; the safe-zone guides cover 65 platform presets.
How do you style and animate captions in After Effects?
Pick a look, then adjust words, motion and keywords. Caption studio has six ready looks (a bold social look, Script hero, Cinematic serif, Stagger stack, Minimal kinetic and Podcast clean), caption templates such as Neon glow and Simple background, five layouts (inline, caption stagger, big word, duo, font pairing) and five font pairings for headline and body text.
- 13 word animations: Pop, Whip, Scale, Bounce, Elastic, Drop, Slide up, Slide left, Slide right, Rotate, Blur in, Fade and Typewriter, with speed, intensity, word gap and placement.
- Word motion on one word, a group or all words: slide, bounce, pop or fade, with an optional shake, jitter, motion blur, flicker or blur.
- Keyword highlighting in a different colour and size; there's automatic keyword detection too.
- Styling controls for font, normal and keyword sizes and colours, box colour and opacity, and padding.
Go easy on keywords. One every five or six words lands; every other word and nothing stands out. Check caption colours against the footage with the contrast checker in Color desk, and for titles beyond captions use the text animation presets.
How accurate are auto captions, and what still needs a human?
Good enough to save the typing, never good enough to skip the read-through. Speech models stumble in predictable places:
- names, brands and product names, especially new ones;
- numbers, prices and dates;
- slang and words borrowed from another language;
- two people talking over each other;
- music sitting loud under the voice.
Read every line once before you build. Fixing a chip takes a double-click; fixing a built and animated caption takes far longer.
How does Caption studio compare with other After Effects caption plugins?
AutoCaption on aescripts is the closest stand-alone option; mouzeclick folds captions into a larger workspace. As shown on 31 October 2026:
| Item | mouzeclick Caption studio | AutoCaption (aescripts) |
|---|---|---|
| Price shown | One plan: $25 a month, $249 a year or $499 lifetime | $8 a month, $80 a year or $150 perpetual |
| Trial | 3 days, no card | 7 days, no credit card |
| Where transcription runs | On your computer | Online through its servers, or on your machine when offline |
| Languages | 100 plus Auto Detect, Hinglish included | 100+ with auto-detect |
| Styles | Six looks, five layouts, 13 word animations | 60+ editable caption styles |
| After Effects | Panel files support CC 2018 or later | CC 2019 or later; list shows 2020 to 2026 |
| Computers | Windows | Not listed on the product page |
Where AutoCaption is the better pick: you only want captions, you'd rather pay once, or you caption Arabic, Hebrew, Chinese, Japanese or Korean, where it offers script-matched styles. AEJuice also lists Auto Captions at $19 a month on its home page. If you also want silence cuts, text presets, maps and colour tools in the same panel, that's the case for mouzeclick; see the AEJuice comparison and the all-in-one After Effects extension guide.
Do auto captions in After Effects work on Mac with mouzeclick?
No. Caption studio is part of the mouzeclick workspace, which is part of mouzeclick on Windows, and mouzeclick's auto-captions are Windows only for now. The Mac page lists what the Mac app does include. On a Mac, Premiere Pro's own Speech to Text or a plugin is the way; check its page for macOS support.
Why are my auto captions wrong, slow or in the wrong place?
- Wrong language. Auto Detect guessed from a noisy start. Set the language yourself, or choose Hinglish for mixed Hindi and English.
- The first run is slow. Caption studio first sets up its local engine (the setup step installs Python and Whisper) and downloads the model once. After that it starts from your drive.
- Captions for footage you trimmed away. Set the range to the layer's span only, or to the work area.
- Lines too long on a phone. Lower words per line and build again; it's faster than splitting layers by hand.
- Everything highlighted. Turn off keywords you didn't choose; automatic detection is a starting point.
Caption your next After Effects comp on your own computer. One plan with everything included, $25 a month, or $249 for a year.
Questions people ask
Does my audio leave my computer?
No. Caption studio transcribes on your computer, and the dialog says no data leaves your machine. The speech model downloads once. Separately, the mouzeclick app needs an internet connection while it runs, for its licence, so keep it online even though the audio stays local.
Can Caption studio make plain subtitles instead of styled captions?
Yes. Next to the styled caption layers, it can build standard subtitle layers from an SRT it writes for you, or export the captions as a JSON file. Plain subtitle layers suit long-form videos and client cuts, and the JSON export is handy when a developer or another tool needs the word timings.
Can I bring captions I already have into After Effects?
Yes. The SRT importer in Motion lab (under Fx presets) brings an SRT file in as text layers in the active comp. Its Subtitle editor then finds and replaces text, changes case and spellchecks subtitle layers in bulk, across the selected layers or the whole comp.
Can it caption Hinglish in After Effects?
Yes. Hinglish is one of the 100 languages, and Hindi can be written in Roman letters the way people type it. Pick it yourself rather than relying on Auto Detect, since mixed Hindi and English speech confuses detection, then read the spellings through once before you build.
Can Caption studio translate captions?
Into English, yes, offline, with the larger speech models. It doesn't translate into other languages. If you need captions in several languages, transcribe once, translate the text elsewhere and bring each language back as an SRT through the SRT importer.
Can I caption several clips at once?
Yes. Choose every media layer in the comp as the source instead of a single layer. Do it once the edit is locked, so no clip moves after its captions are built, and read through each clip's words before you build, as speakers and audio quality usually differ between clips.
How much do auto captions in After Effects cost with mouzeclick?
Caption studio is included in the one mouzeclick plan: $25 a month, $249 a year or $499 lifetime, with every other tool and extension. Try it free for 3 days with no card. See pricing; a creator or friend code takes 10% off a new customer's first 30-day or 1-year payment.
Sources
Checked on 2026-10-31. Prices and features change: each link goes to the maker's own page.