Paste your video. Get a thumbnail that actually matches it.
Drop in a YouTube URL or paste your transcript and FatThumb's AI reads what the video is about — then generates 1–4 thumbnail variations featuring your own face, sized exactly for YouTube Studio.
How it works
Three steps from zero to upload-ready.
Paste a YouTube URL or your transcript
Drop in a link to any public YouTube video and FatThumb fetches its captions automatically. No published video yet? Paste at least 100 characters of your transcript or script instead.
The AI analyzes your content
FatThumb builds a structured read of the video: a summary, the target audience, an exaggerated story angle for the visual hook, and a concrete thumbnail concept derived from all three.
Generate with your face locked in
Attach your Person profile and generate. You get 1–4 variations built from the content analysis, each showing your exact face, each an exact 1280×720 PNG ready for YouTube Studio.
Features
Everything you need for the click.
From link to thumbnail concept
Paste the URL of any public YouTube video with captions. FatThumb fetches the captions and reads what the video actually says — the thumbnail is built from your content, not from a generic prompt.
Thumbnails before you film
Paste 100+ characters of your script or transcript and generate the thumbnail before the video exists. Plan the visual hook first, then film toward it.
Summary, audience, story, concept
The analysis extracts a summary of the video, who it is for, an exaggerated story angle — the dramatic framing thumbnails need — and a visual concept that ties them together.
Your face in every result
Pair the flow with a Person profile and every generated variation uses your exact likeness. The content drives the scene; your face reference stays the constant.
1–4 directions per video
Each run produces up to four variations from the same content analysis. Compare them side by side, star the strongest, and keep the rest in version history.
Exact 1280×720 PNG output
Every download is an exact-spec 1280×720 PNG. No cropping, no resizing, no retouching — upload it straight into YouTube Studio.
FAQ
Questions & answers
How does the YouTube URL option work?
You paste a link to a public YouTube video and FatThumb fetches the video's captions. The AI analyzes that text — producing a summary, target audience, an exaggerated story angle, and a visual concept — and uses it to generate thumbnails that match the content. The analysis is text-based; FatThumb does not pull frames out of the video.
Can I upload a video file?
No. FatThumb works from a YouTube URL or pasted text, not from video files, and it does not extract frames from footage. Instead of grabbing a frame, it generates a purpose-built thumbnail image from your content analysis and Person profile — a composed 1280×720 image rather than a screenshot.
What if my video has no captions?
Paste the transcript or script text directly instead — anything from 100 characters up works. The AI runs the same content analysis on your pasted text that it would run on fetched captions.
Can I make the thumbnail before the video is published?
Yes. Paste at least 100 characters of your script or outline and generate from that. Many creators design the thumbnail first and shape the video's hook around it.
Will the thumbnails show my actual face?
Yes, if you attach a Person profile. Create one once by uploading 1–5 photos of your face, then select it when generating — every variation uses that exact likeness. Without a Person profile, the AI invents a face from the content description.
What format and resolution do I get?
Every thumbnail is an exact 1280×720 PNG, the standard YouTube size. Generation uses credits at your plan's per-thumbnail rate; the free plan includes 5 watermarked thumbnails so you can try the flow before upgrading.
Your video already knows what its thumbnail should be
Generate your first 5 thumbnails free — no card, no designer, your face consistent from the first run.