AI Study Habits YouTube Thumbnail
See a face-led YouTube thumbnail generated with GPT Image 2.5 Flare, built for a click at feed size from one uploaded reference creative.

Example breakdown
Why it stays on-brand
The charcoal ground, saturated yellow highlight block and white display type all arrive from the reference, none of them named in the prompt.
Marketing use
Use it for video thumbnails, study-skills course promotion, webinar artwork, or education channel covers.
Prompt
A YouTube thumbnail designed to be clicked. Close-up of a young woman in her early twenties occupying the right third of the frame, head and shoulders, turned slightly toward camera with direct eye contact and an expression of genuine startled realisation, eyebrows raised, mouth slightly open. Warm key light on her face from the left, background falling off to near-black on the left side. Behind her, softly out of focus, a dark chalkboard covered in faint handwritten equations. On the left two thirds, a three-line headline in heavy condensed uppercase reading "THE WAY" / "YOU STUDY" / "IS WRONG", with the words "IS WRONG" sitting on a solid saturated yellow block. Beneath it, one short handwritten script line reading "here's the fix" with a hand-drawn brush underline. Photographic realism for the person and the room. Crisp flat 2D graphics for all typography. Hard contrast, punchy, legible when shown small. No logo, no watermark.
How this was generated
Brand DNA came from your reference visual. That DNA was applied as rules on top of the prompt used for this asset (shown in the Prompt section above).
- 1Reference creativeBrand DNA source
Upload path: BrandGen reads palette, typography, and style cues from one reference asset.

- 2Apply DNA to the prompt
Every generation runs through that DNA as hard constraints (colors, type, visual style, tone). The creative intent is entirely in the prompt already shown above.
- 3Final outputEducation
The same wording with different DNA would look like a different brand. This image used the Brand DNA from step 1.




