Burnin transcribes the speech in your video right on your iPhone, no network, no account, and burns editable, auto-generated captions into a new file you export. Nothing about the video or its audio ever leaves your phone.
No account, no ads, no subscription. Speech is transcribed on the phone itself, so once the language pack is on it, captioning keeps working with no signal.
Most caption apps run in the cloud because that is the easier engineering path, not because your video needs it to. Burnin does the same job the other way around.
| A typical cloud captioning app | Burnin | |
|---|---|---|
| Where your video goes | Uploaded to a server to be transcribed | Stays on your phone, transcribed on-device |
| Account | Sign-up required to export | None, ever |
| Internet connection | Needed every time you transcribe or export | Only once, to download a language model |
| Price | Often a subscription for full-length or HD export | No subscription, no ads |
A video shot for one place rarely fits another. Burnin crops and exports for the shape a platform actually wants, with captions burned in to match.
The source video's own aspect ratio, no crop applied.
Full-height vertical crop for short-form video.
A square crop for feed posts.
Standard landscape, for everywhere else.
Higher resolution means a bigger file and a longer export, and Burnin shows you that trade before you commit to it rather than after the export has already started.
Import, transcribe, edit, export. Everything below works with no account.
Speech is turned into text using Apple's on-device speech transcription, not a server somewhere. Once the language pack is downloaded, captioning keeps working on a plane, underground, or with the data switched off.
Burnin detects the spoken language on its own, with a manual picker for when you would rather choose. Per-language packs download once, then everything runs fully offline.
A raw transcript is automatically segmented into caption-length lines, so what appears on screen is something you can actually read at a glance, not a wall of text.
Burnin notices spans that likely got dropped during transcription, but only retries one when an independent speaker-diarization pass confirms real speech actually overlaps it, so it never invents captions over silence.
An on-device speaker-diarization model works out who spoke when. Captions get a "- " prefix at every speaker change, so a multi-speaker video is labeled with no manual work.
Tap a caption or the add button to open a full-height editor sheet, drag its start or end on a visual timeline strip, and drag the playhead itself to scrub the video. Neighbor-clamping keeps captions from overlapping as you trim.
Every edit is undoable, tens of steps deep, and your edits are saved automatically as you make them, so closing the app mid-edit loses nothing.
The video preview shows captions burned in exactly as they will appear in the exported file, updated as you edit.
Before you share an export, Burnin offers a title and description generated on-device from the transcript, in the tone you pick, with an on-device fallback on iPhones without Apple Intelligence.
Every video you import becomes a project you can rename, delete, or select many of at once. Projects live on your phone, with no cloud sync and no account.
If the original video file goes missing locally, Burnin can recover it from your Photos library rather than losing the project.
Importing, transcribing, editing and exporting all happen on your phone, and every project stays there. There is no sign-in, so there is nothing to sync and nowhere for a video to go but your own device.
| Importing a video | Nothing. The video is read from Photos or Files and stays on the phone. |
| Downloading a language pack | A request for that language's model file only. No account, no identifier. |
| Downloading the speaker model | A one-time request for the diarization model, around 100MB, then it runs offline. |
| Transcribing, diarizing, editing | Nothing. All of it runs on the phone. |
| Exporting a video | Nothing. The captioned file is written to your phone. |
On-device transcription needs a fairly recent Neural Engine, so Burnin supports iPhone 12 and later. Older devices, and the Simulator, are not supported. Burnin runs on iPhone and iPad, though the app is not iPad-optimized yet. An older iPhone can still open a project to write captions by hand or bring in a ready-made subtitle file; it just can't transcribe speech itself.
Sending video to a server would mean an account, a connection, and someone else's copy of your footage sitting somewhere. Running transcription and diarization on the phone means none of that has to exist. It costs a one-time model download and it means the oldest supported iPhones do the work themselves rather than a cloud doing it for them.
For iPhone, on the App Store. No account needed to use it.
No. Importing, transcribing, editing and exporting all work with no sign-in of any kind, and there is no account to create in the first place.
Only once, to download the language pack for the speech you are transcribing and, if you use speaker labeling, the diarization model. After that, transcription, gap detection, diarization, editing and export all run with no connection at all.
MP4 and MOV, imported from your Photos library or the Files app, up to 10 minutes long.
No transcription is perfect, and Burnin's is no exception. The in-app editor exists because every caption is meant to be checked and adjusted, not accepted blindly. How it works is candid about what's still approximate.
No. Burnin is an iOS app only, with no Android version, no Mac app, and no Apple Watch app.
There is no subscription and there are no ads. Every export resolution currently works with no paywall. A future one-time purchase for something added is possible, but nothing is locked behind a subscription.
Burnin