Claude artifacts and ChatGPT canvases are great for getting an idea onto the screen fast. The catch: what you get is code. You can't hand it to a designer, put it on your brand or publish it without a developer.
With Vev connected, that's one more message. Your assistant sends the prototype to Vev, and Vev rebuilds it as native Vev sections: real frames, text, images and icons that you edit, restyle and publish like anything else you build in Vev. It isn't a screenshot or an embedded iframe, and there's no code to maintain.
Before you start
Connect Vev to your assistant once. See Connect Vev to Claude, ChatGPT and other AI tools (MCP).
The workflow
Prototype. Ask for what you need and iterate until you like it: "Make a landing page for our new running shoe: bold hero, three feature cards, testimonials and a pricing table. Dark theme, lime accent."
Send it to Vev. "Import this into Vev, in my project Spring Campaign, on a new page called Running Shoe." No project yet? "Create a new Vev project for it."
Wait a few minutes. Each section appears on the page right away as a placeholder and fills in when its import finishes, usually in 2–5 minutes. You can keep working in the editor meanwhile.
Make it yours. Open the page in Vev (the assistant gives you the link), swap in brand fonts and colours, adjust spacing for each device, add animations and publish.
Prompts that work well
"Import each section of this page into Vev as its own section, in order." Separate sections are easier to edit, reorder and reuse than one long one.
"Only import the pricing section into my Product Launch project, above the footer."
"Check on the import and tell me the match score and anything that didn't carry over."
"The testimonials came out at 72. Fix the prototype so it's simpler, then import that section again."
What Vev does with it
Renders your prototype in a real browser at desktop and mobile width, so React and Tailwind artifacts work as they are.
Rebuilds it natively with AI, using exact measured sizes, colours and fonts, with separate desktop and mobile layouts.
Checks its own work. A vision model compares the Vev section with your prototype side by side, and Vev makes up to three rounds of fixes. You get a match score from 0 to 100.
Makes repeated things components. Cards, plans and team members become one main component with instances, and a highlighted card becomes a variant. Edit the component once and every copy follows.
Keeps hover effects as component states, and turns interactive parts (tabs, accordions, sliders, forms) into widgets built by Element AI.
Copies your assets. Images, SVG logos and fonts are copied into the project, so nothing points back at the prototype.
Tips for the best result
Use real images. Ask the assistant to use public image URLs (for example from your website or Unsplash) instead of placeholders. Grey boxes import as grey boxes.
Design for mobile too. Ask for a responsive layout. Vev reads the mobile version for its mobile layout.
Keep the copy final-ish. Text comes across word for word, so it's easier to fix it in the chat first.
One section per import. A page made of clear sections imports better than one huge block.
Good to know
Each section import uses AI credits from the project's team, about $0.15–0.30 per section. An import doesn't start when the team has less than $1 of credits.
Scroll animations and custom scripts aren't carried over. Add animations in Vev instead.
Videos come in as a still image. Replace them with the real video in the editor.
Content hidden behind an interaction, such as a closed menu, is left out unless it's inside a widget.
Also see: Import a section from any website.
