Captions written on the iPhone, exported as SRT or burned in

How the transcript is made

The Captions tool in the editor uses Apple’s on-device speech models. The first run may download a language model; after that the transcription runs on the phone and nothing is sent anywhere. The transcript is split into short cues that end at the end of a sentence, so the words on screen read the way people speak instead of running on in one block.

Correct the words, not the take

Any line can be edited by hand. A corrected line shows in full white so you can see at a glance what you changed. A stumble that would once have meant another take becomes a line you fix in the transcript, which is also why the teleprompter and the captions belong together: read it once, correct it once.

Two ways out

• SRT file. Exported through the iPhone share sheet, because the Photos app does not keep caption files. Upload it beside the video on the platforms that support captions, and viewers can switch them on and off.
• Burned in. The words are drawn into the picture at export. It is permanent, so it is off by default and you choose it per export. In a split layout the caption band sits in the main camera’s half rather than across the seam.

Free, with one honest exception

Transcribing, correcting and preparing the SRT are free. Exporting the video with the words burned in is a video export, and video export is what the PRO subscription unlocks. An SRT is not a video, so it leaves the app without PRO.

Which iPhones

Captions need iOS 26, where Apple’s on-device speech models became available to apps. The app checks that the phone can actually transcribe rather than trusting the iOS version alone, and a phone that cannot is simply not offered the tool; there is no greyed-out button to wonder about. Everything else in Dual Recorder works from iOS 15.