Turn a screenshot or Figma design into working code

+ Produces editable designs you can paste into Figma as well as HTML/CSS.
− Experimental Labs product with generation caps and limited iteration.

+ Supabase auth, database, and edge functions are wired up without any configuration.
− Credits run out fast when debugging, since every fix attempt costs a message.

+ Long-standing Figma plugin that respects auto-layout and produces responsive code.
− Generated components still need cleanup before they are production-ready.

+ Iterative loop: annotate the generated output on the canvas and regenerate.
− It is a demo of the tldraw SDK rather than a maintained product.

+ Simple, widely used, and works with GPT, Claude, or Gemini vision models you supply.
− Produces one-off static code with no design-system awareness.

+ Mitosis-based output targets React, Vue, Svelte, Angular, and more from one design.
− Best results require Builder's design-system mapping setup first.

+ Runs npm, dev servers, and builds entirely in the browser so nothing is installed.
− Token-metered pricing burns through the monthly allowance on larger projects.

+ Lightning mode tags and converts an entire design in one pass with minimal manual tagging.
− Output structure can be verbose and hard to maintain in larger apps.

+ Output uses shadcn/ui and Next.js idiomatically, so it drops straight into real projects.
− Heavily biased to the Vercel stack; other frameworks get worse results.

+ Starts from your existing Figma frames and design system, not a blank prompt.
− Output is prototype-grade and harder to hand off to a real codebase than v0.