Agentically creating Sanity Studio thumbnails

Want AI-generated thumbnails for your Sanity page builder, in one prompt? We walk through the skill we use on top of Turbo Start Sanity, our open-source Next.js + Sanity starter. Try it yourself or…

Frequently asked questions

What is the Sanity Studio thumbnails skill?
It's an agent skill we run on top of Turbo Start Sanity that generates and assigns custom preview images for every page-builder block in Sanity Studio. You drop a single prompt, the agent works through your schema, and you end up with a complete thumbnail set instead of editors picking blocks from a wall of identical icons.
What is Turbo Start Sanity?
Turbo Start Sanity is Roboto Studio's open-source starter for building production websites on Next.js and Sanity. It ships with the page-builder patterns, schema definitions, and visual editing setup we use on every client project, so you don't start from a blank Studio.
How long does it take to generate Sanity page builder thumbnails?
In the demo it took about five to six minutes end to end, depending on connection speed and how busy the model provider is at the time. The skill runs unattended once kicked off, so you drop the prompt, walk away, and come back to a finished set of preview images.
Why bother with thumbnails for a Sanity page builder?
Editors pick blocks faster when they can see what each one renders as, instead of reading a label. Custom previews are part of why Sanity's page builder feels editor-friendly in the first place, and skipping them is one of the quickest ways to make a Studio feel generic.
Can the agent regenerate thumbnails I already have?
Yes. The prompt has an option to delete existing thumbnails before it runs, so you can rerun the skill any time the design language changes without ending up with a mix of old and new previews.
Do I need Turbo Start Sanity, or will this work with my own setup?
The skill is built around the conventions Turbo Start Sanity uses, so it works out of the box there. On a custom Studio you can still use it, but you'll likely need to adapt the prompt to match how your blocks are named and where preview images get stored.
What is a Sanity AI agent?
In this context, an AI agent is a long-running process driven by a prompt and a set of tools — read the schema, generate an image, upload it, write the reference back to the document. It's how we automate jobs in Sanity that would otherwise need a developer for an afternoon, like generating thumbnails for every page-builder block.

Continue watching

Build Sanity CMS Pages Faster with MCP
765

Build Sanity CMS Pages Faster with MCP

Build full Sanity pages in minutes using the Sanity MCP in Claude. I walk through the setup, prompt Claude to generate a services page with zero copy, and review how well it picks page builder blocks.

Watch video
Running git work-trees with Portless & Conductor
1221
0:42

Running git work-trees with Portless & Conductor

Here's how we set up Portless with Conductor so each of your git worktrees gets its own isolated environment. The setup is pretty straightforward: you configure Portless once, point it at your code…

Watch video
Your first AI agent
1239
9:08

Your first AI agent

Looking to build your first AI agent workflow? We share the stack that powers our current lead enrichment, and the open source Lead Agent repo from Vercel. Go build it, or ask us to build it for you.

Watch video

Get in touch

Fill out the form below and we'll get back to you