Camera mode (called
Capture internally) is the second option in the
Create your video mode sheet (see
Record your voice for how to reach that sheet). Unlike Voice, which only records audio, Capture records video of
you — and uses that footage as a motion reference to drive the AI character that ends up in your final video.
HEADS-UP
Your video isn't published — it's a motion reference
Capture is not "post a selfie video." Your recording is a private motion source used to clone your facial expressions and hand gestures onto the AI character. The original footage is discarded after the clone is built. The published video shows the AI character moving the way you moved.
1. The Camera Explanation sheet
Tapping Camera (with Capture access) doesn't open the camera immediately. A short bottom sheet appears first to set expectations.
Be expressive — we'll clone your face & hands
Video won't be saved — only used for cloning
Three bullets explaining what Capture actually does — then the Start CTA with your plan's max duration.
The sheet has the title "Record video." and three bullets:
Record your voice — the camera takes audio too; you don't need a separate voice pass.
Be expressive — we'll clone your face & hands — gestures and expressions matter; static delivery wastes the mode.
Video won't be saved — only used for cloning — the privacy guarantee. Your footage doesn't end up in your feed, your library, or anywhere else.
At the bottom, a gradient CTA labelled Start — 30s max. Tap it to dismiss the sheet and open the full-screen recorder.
2. The camera screen — idle state
Walking through the venue, pointing out where the ceremony will be —
Idle state: 'Your expression.' title, teleprompter overlay, dashed silhouette guide, gradient record button at the bottom.
Full-screen camera view with four overlays stacked on top:
The "Your expression." title
Top-left of the screen, in white. Replaces what would normally be a scene name. It's the screen's voice cue — a reminder that howyou deliver matters more than what's behind you. Disappears once recording starts.
The teleprompter
Same component as the Voice screen, but with a translucent dark background card so it's readable against any camera scene. Scrolls word by word at your saved WPM setting (default 150, range 80–240). Adjust with the slider; the value persists across sessions.
The silhouette guide
A dashed head + shoulders outline rendered in low-opacity white, with the caption "Step back. Face forward." Helps you frame yourself before recording. Disappears once recording starts.
The record button
Round, with the brand cool-gradient and a video-camera icon — the screen's signature brand moment. Below it: "Tap to record" hint. The whole bottom area sits over a subtle dark gradient so white text remains readable.
"or — Choose from library"
Below the record button. Lets you skip filming and pick an existing video from your phone instead. Promoat validates the duration against your plan limit and rejects videos that are too long with a clear alert.
3. Countdown and recording
Tapping the record button triggers the same 3-second countdown as the Voice screen — a large gradient "3 / 2 / 1" with a "Get ready..." caption. The teleprompter pauses during the countdown and starts scrolling the moment recording begins.
and don't miss the rooftop terrace —
Recording state: timer pill at the top of the controls, red stop button, teleprompter still scrolling.
While recording, the controls shift:
A pill at the top reads "0:08 / 0:30" — red dot for the live indicator, elapsed / max in tabular numerals.
The gradient record button becomes a red square (stop). Tap to end early.
The "Your expression." title and silhouette guide disappear so the frame is clean.
Recording auto-stops at the 30-second cap. When the camera stops, Promoat immediately uploads the file and you'll see a "Preparing clip…" overlay during the upload step.
TIP
Light, distance, framing
The silhouette guide is calibrated for the cleanest motion-clone result: face front-lit, framed waist-up, two-arm-lengths from the phone. Diffuse window light works well; harsh overhead light gives the cloner less to work with.
4. Uploading instead of filming
Tapping Choose from library opens the iOS / Android media picker filtered to videos. Pick any clip up to 30 seconds long. Promoat treats it the same as a fresh recording — the audio is extracted, the motion is extracted, and the original is discarded after the clone is built.
HEADS-UP
Duration validation
Videos longer than 30 seconds are rejected with an
"Audio Too Long"
-style alert. Trim outside the app first.
5. After the take
Once the upload finishes, the screen returns to the Create Video summary with your clip attached. Tap Use this take to advance to the Caption options sheet. From there the flow is identical to a Voice take — you pick caption style, confirm cost, and trigger the render.
Camera takes run a per-scene motion-clone job in addition to the lipsync, so credit cost is materially higher than a plain Voice render. The exact number is shown in the caption-options sheet before you commit.
Tap Record on a Ready idea, then Camera in the mode sheet
Read the three-bullet explanation, tap Start
Frame yourself with the silhouette guide
Tap the gradient record button — wait through the 3-2-1 countdown
Read the teleprompter while gesturing naturally
Tap the red stop or let it auto-stop at 30s
Wait for 'Preparing clip…' to finish, then Use this take