Thumbnails made from your best frame

The AI pick

Up to eight candidate frames from across the footage are compared, along with your brief, the transcript and your reference videos’ thumbnails, both the most viewed and the least viewed. It picks the frame with the strongest expression, eyes and product in shot, says why, and writes three hooks of two to four words that stay within the claims your brief says to avoid.

When you have reference thumbnails, the title’s font, colors, capitals and box are matched to your best performers.

Built to read at feed size

Thumbnails are seen small, in a feed, so every render is set up to read at that size.

• The crop punches in so your face fills the frame, as far as the footage’s resolution allows
• Words sit on the side your face leaves free and stay off your eyes and mouth
• Titles start big, and words you wrap in *stars* take an accent color
• A little extra contrast and sharpness, with hue and saturation as filmed
• An optional light touch-up that smooths brightness only, so makeup and product colors stay exact

One design, every format

The thumbnail is anchored to a moment in your footage, so the YouTube video, the Short and the Reel all show the same shot, each cropped to its own size. Titles wrap and shrink to stay inside each platform’s safe area, and YouTube posts from the content calendar carry the custom thumbnail.

You stay in charge

Pick a different frame yourself, rewrite the words, or switch to another suggested hook with one click.

Questions