YouTube / video thumbnail

A click-driving thumbnail or cover for a video on YouTube, Shorts, Reels or TikTok. Packaging for something else rather than an ad in its own right.

When to use it

Any video you publish. The thumbnail is the single biggest lever on click-through rate for YouTube, and it is the most commonly neglected asset in a video workflow.

The thing to get right

There is a dedicated thumbnail pipeline behind this (frameworks, emotion casting, identity locking and multiple takes), reachable as its own tool rather than through the normal ad flow.

The format at a glance

• Output: a still image ad.
• Structure: a single frame, with the copy composited on top.
• Family: Production assets.
• Recipe id: yt_thumbnail, which is what you name when you ask for it by hand or from an agent.

Is this the same as the thumbnail tool?

There is a dedicated pipeline behind it, with frameworks, emotion casting, identity locking and several takes to choose from, reachable as its own tool rather than through the normal ad flow. Use this recipe when a thumbnail is one part of a larger set you are making together.

Related formats

Things that feed an ad without being one: a video thumbnail, a locked character reference sheet, a restyled interior.

• Character / model sheet: a multi-view character reference sheet (full body and close-up, turnaround or expression grid) of ONE original person, the identity lock other creatives reference
• Interior / real-estate restyle: a listing-grade restyle of an interior or property photo (virtual staging, modern renovation, declutter or day-to-twilight) that keeps the SAME room, improved

More still image formats: Studio product shot, Checklist / save this, Lifestyle / in-use, Tweet / X post.

The full catalog is on the creative recipes index.

Models that render it

This format is rendered by an image model, with the copy composited on top afterwards. GPT Image 2.5 is the default pick. The image models page compares the rest, including Grok Imagine, Bria HD and Krea 2 Turbo.

Hermoso plans and renders a youtube / video thumbnail from your brand rather than from a prompt you write, composites every on-screen word after the render so the copy comes out exact, and exposes the same format to an AI agent over MCP. How Hermoso renders a format covers all three.