I make CapiCap on my own. These answers are about version 1.8, the one on the App Store now. For the whole process, from picking a video to saving the captioned one, there's a step-by-step guide.
Cost
Is CapiCap free?
Up to a point. It's free to download, with every feature, and 3 exports every calendar month are free. Past that, Lifetime is a one-time purchase ($9.99 in the US, priced locally elsewhere) that removes the limit for good. There's no subscription, and there are no ads. Your 3 come back at the start of each month, in your device's time zone; once they're used, the Lifetime offer shows the date the next 3 arrive.
What counts as an export?
A video render that finishes, counted the moment it does, even if it's the same project again. A render that fails or that you cancel costs nothing. Saving or sharing a rendered video again, and taking its .srt, is free as often as you like. Export .srt never uses one of your 3, before or after rendering; like saving and sharing, it only locks once the month's 3 are gone and you don't have Lifetime.
I bought Lifetime. How do I get it on another device?
Open CapiCap on it while signed in with the same Apple Account. The app checks your purchase with the App Store when it starts, so Lifetime normally turns on by itself. If it doesn't, tap Restore Purchase: it's on the Lifetime offer, and in About CapiCap (the info button on Home) once you've reached the monthly limit.
Your video
Does CapiCap upload my video?
No. Transcription, editing and export all happen on your device, and your videos, audio, transcripts and exports are never uploaded, to me or anyone else. The app goes online once, on first launch, to download its speech model; after that it works with no connection at all, in airplane mode too. There's no account, and no analytics, tracking or advertising SDKs. The privacy policy has the details.
Is there a watermark?
No, not on free exports and not with Lifetime. The video you export carries your captions and nothing else.
Does the export keep the original quality?
Yes, with one exception. It keeps the video's resolution, frame rate and bitrate, 4K60 included, and copies every audio track as it is. The exception is HDR: an HDR video exports as SDR (Rec.709). The editor previews that same SDR picture, so the export matches what you saw.
Can I export an .srt file?
Yes. Export .srt sits under Export video and gives you a standard SubRip file with each caption's text and timing, for YouTube, Vimeo or a video editor. You don't have to render the video first. There's a guide to making one.
Making captions
How long does transcription take?
On recent devices, less time than the video lasts, and it's quickest on iPhone 14 Pro or newer, iPads with an M-series chip and Macs. On an M1 Pro MacBook Pro, 62 minutes of audio took about 11 minutes. The first transcription after installing is slower, because the speech model is being prepared (more on that below).
How accurate are the captions?
A lot depends on the recording. Clear speech close to the mic needs the fewest fixes; music, people talking over each other, wind and heavy accents cause most of the mistakes. On iPhone 13 and later, iPhone SE (3rd generation), iPad mini (6th generation) and later, iPad (A16), every iPad with an M-series chip and every Mac, CapiCap runs OpenAI's Whisper Large v3, the most accurate speech model it ships. The app picks the model itself; About CapiCap (the info button on Home) shows which one yours uses. Anything it gets wrong, you can fix before you export.
Can I fix the words it gets wrong?
Yes. The whole transcript is a list of captions synced to the video. Tap a caption to edit it; split, merge, add or delete captions; move a caption's timing on the scrubber strip; and undo or redo anything. Projects save themselves as you go.
What can I change about how the captions look?
You start from one of twelve looks (Classic, Impact, Karaoke, Glide, Chip, Cover, Outlined, Hype, Pop, Boxed, Editorial and Typewriter), and Customize changes the font, weight, capitals, size, outline, colors, shadow, background, motion and position. You can save the result as your own look or make it the default for new videos. Lines, on the Layout page, sets one, two or three lines per caption, two by default, and captions break at natural phrase boundaries. The preview is drawn by the same renderer as the export, so the video comes out the way it looked in the editor.
Can I put the captions behind a person?
Yes, with the Cover look: the caption sits behind the person in the shot, like a headline on a magazine cover. Apple's Vision subject detection finds the person on the device, frame by frame, and Place behind the person finds a height where that works. Where the person would hide too much of a word, the caption comes forward so it stays readable. The editor shows the effect on a paused frame. I wrote up how it works.
Devices and languages
Does it work on iPad, Mac or Android?
iPad and Mac, yes; Android, no. CapiCap is one app for iPhone and iPad with iOS 17 or later, and for Macs with Apple silicon (M1 or later) on macOS 14 or later, where it runs as the iPad app in a resizable window. On the Mac, and on iPad in landscape, the preview and the whole transcript sit side by side. There's no Windows or web version either. More on captioning on a Mac.
Which languages does it understand?
99+, the languages OpenAI's Whisper knows, from English, Spanish, Hindi and Arabic to Welsh, Swahili and Tagalog. It detects the spoken language by itself, or you can pick it. Here's the full list.
Can it translate the captions?
No. Captions are always in the language spoken in the video. If you need a translation, export the .srt and translate it with another tool.
Can it label who's speaking?
No, it doesn't detect speakers. If a caption needs a name, type it into the caption.
When something goes wrong
Why is the first launch so slow?
On first launch CapiCap downloads its speech model, once, so it needs a connection then. On devices that run Whisper Large v3 it then prepares the model for the Neural Engine, which can take a few minutes; Home shows an estimate, and the device may get warm. That happens once per app update. On an M1 Pro Mac it took 134 seconds, and later loads took about 2 seconds.
It picked the wrong language. How do I fix it?
While it's transcribing, tap the language button under the progress bar and pick the right one; transcription starts over in it. Later, tap the language button at the bottom of the caption list, pick the language and tap Re-transcribe. That rebuilds the captions from the video, so edits you made to them are lost. Detection listens to up to three 30-second stretches of the video, so a clip that opens with music or mixes languages can fool it.
The speech model won't download.
Check that you're online and tap Retry on Home; the download picks up where it stopped. If there isn't enough free space, CapiCap tells you how much to free up. Once the model is downloaded, it stays through app updates.
Save to Photos doesn't work.
If CapiCap isn't allowed to add videos to your Photos library, the export screen says so and offers Open Settings, where you can allow it. If saving fails for another reason, use Share… instead. Before you export, CapiCap also warns you when there isn't enough free space for the new video.