How to Add Captions to a Video | ReelCraft

The short answer

ReelCraft gives you two routes to captions. The automatic one: start a system-wide screen recording through its Broadcast Upload extension, and the app runs speech recognition over the audio that recording captured, turning what was said into a caption track you can edit. The manual one: add an animated title or subtitle layer and type them yourself. Both come out as ordinary text layers, so both take the same fonts, strokes, shadows, glow and animation — and both are part of the exported file rather than something you bolt on in a second app afterwards.

If the clip is one you already shot with the camera, the automatic route does not apply to it; you type those. What ReelCraft does change is the typing: captions live on a frame-accurate timeline, so you are placing them rather than guessing at them.

Why the searcher is here

"How do I add captions to a video" is rarely a question about a button. It is one of three questions wearing that coat:

• "I can type captions. Why is it taking me forty minutes?" Because typing is not the work — timing is.
• "I cannot type three hundred words of dialogue." So: can something listen for me?
• "My last two attempts were wrong in a way I could not see." Because the caption was added after export, and once it is baked into a file, a two-frame slip is a two-frame slip forever.

The middle one is the interesting one, and it is the one an editor can genuinely answer. Speech recognition ships on every recent iPhone. What is not automatic is routing that recognition into a caption track you can then correct, rather than a finished video you can only accept or reject.

Route one: speak it into a screen recording

ReelCraft ships a Broadcast Upload extension — a system-level screen recorder. Start one, make the recording you need while the words are being said, stop it, and import it into a project. Speech recognition produces a caption track from what it heard.

Three things about that route are worth being concrete about:

1. The input is the recording's audio, not your microphone by default. What gets transcribed is what the broadcast extension captured — the app walkthrough, the tutorial, the video call, the thing being demonstrated. If the words were not in that audio, there is nothing to transcribe.
2. It is a track, not a verdict. The source describes it as a caption track you can then edit, and that is the honest framing. Recognition output is a draft: names, jargon, brand terms and anything said over music are exactly the parts worth checking.
3. It suits a specific kind of video. Tutorials, app walkthroughs, screen-recorded explainers, anything where the words matter more than the picture.

What it is not is a general-purpose transcriber for your camera roll. Anyone telling you a phone editor will "just add captions to any clip" is describing two different features.

Route two: type them, and make the typing bearable

Captions in ReelCraft are animated titles and subtitles. That single fact decides everything else about them:

• Fonts — the full set, not a single default face
• Strokes and shadows — the difference between a caption readable over a busy background and one that disappears into it
• Glow — for captions sitting over dark footage
• Animation — captions can animate in rather than appearing flat

And because they are text layers on a multi-track timeline, they are independent of the video underneath. You can place one, retime it, split it into two captions where the speaker pauses, and trim it — while the clip beneath it keeps its own length. A talking-head clip and its captions are two tracks, not one file.

The playhead is frame-accurate with pinch-to-zoom, which is the part that makes manual timing tolerable. You can zoom to the individual frame where a word starts and place the caption edge on it, instead of nudging by eye and re-previewing to see if it stuck.

Timing: why on-timeline beats after-export

This is worth stating plainly, because it is the actual difference between editors.

A caption added after export has to be re-timed by eye against the finished video. It is a guess loop: play, pause, guess, re-render, check. And because the guess is now baked in, a caption that is late is late for good.

A caption on the timeline is a layer you can move and check at any frame while you are still editing. Retiming costs one frame of playhead work, not one export cycle. On a clip where the speaker changes pace — which is most clips — that difference is the difference between ten minutes and an hour.

What comes out of the export

Captions are layers on the timeline, so they are part of the exported file rather than a separate step. You do not export, open another app, and line text up against your own video.

The rest of the export path:

• Up to 4K at 60fps, no watermark with ReelCraft Pro, saved straight to your Photos library
• Editing and exporting are free to use; Pro unlocks high-resolution, watermark-free export plus premium templates and effects
• Pro is an auto-renewing subscription billed through your Apple ID, manageable in Settings → Apple ID → Subscriptions. Figures are per-storefront and move, so check the live price in the app rather than a figure written down anywhere.
• One honest check to make yourself: captions here are timeline text layers inside the export. If your destination platform expects a separate subtitle file, verify that in your own export before you upload rather than assuming.

Where the automatic route stops being the right answer

• Camera footage. Recorded clips are not screen recordings. Type those.
• Music over speech. Recognition competing with a music bed is where transcription quality degrades most; the draft will need more correction than it saves.
• Non-speech text on screen. A UI walkthrough's on-screen labels are not transcribed — captions are built from speech.
• Nothing captured means nothing generated. If the recording missed the audio, there is no track to produce, and the manual route is the only one left.

For those four cases the timeline is not a consolation prize. It is still the difference between placing captions and guessing at them.

What ReelCraft does differently

ReelCraft runs entirely on device. Editing, rendering and export happen locally on Apple's AVFoundation and Metal frameworks — footage, projects and exports stay on your phone and are never uploaded by default, and there is no account required. For captions specifically, that means speech recognition runs on the audio you already have locally, and nothing you film leaves the device to become a caption track. It also means the whole workflow works with no connection at all.

Two honest limits. First, the automatic route is genuinely narrow — it is a screen-recording feature, not a universal transcriber, and pretending otherwise would waste your time. Second, no editor removes the need to check recognition output; ReelCraft makes the correction cheap by making the track editable, which is as far as that goes.

If your captions have to be right rather than merely present, the combination to look for is the one ReelCraft has: speech recognition as a starting point, a frame-accurate timeline for the corrections, styled text layers for the finishing, and no second app in between. ReelCraft is on the App Store — it needs iOS 17 or later, and you can decide whether it fits from the product page first.