Screenshot-to-App
A generation mode where the user drops one or more screenshots of an existing UI and the AI rebuilds it as a working application. VULK pairs vision models with the brand engine to match colors, fonts, and layout.
Screenshot-to-App
Screenshot-to-app turns a static image — a Dribbble shot, a competitor's homepage, a hand-drawn sketch — into a live, editable codebase. A vision-language model extracts the visual hierarchy (sections, components, copy, fonts, colors), the brand engine maps the palette and typography to the closest production-ready tokens, and the generation agent emits the corresponding React or Next.js project.
In VULK, screenshot-to-app is invoked by dragging an image into the chat composer or pasting from clipboard. The image is sent to a multimodal model (Gemini 3.1 Pro, Claude, or GPT-4o depending on the task), the output is a structured spec, and the spec is passed to the same intent modeler used for text prompts. The brand engine then locks the color palette and font pairing — typically within 90 seconds the live preview matches the screenshot at the section level.
Video-to-App
A generation flow where the user uploads a video — a Loom walk-through, a competitor demo, a Figma prototype recording — and the AI builder reconstructs it as a working app. VULK extracts frames, runs vision models, and synthesizes the codebase.
URL Clone (AI)
A generation flow where the user pastes a public URL and the AI builder fetches, analyzes, and rebuilds the site as editable source code. VULK uses headless Chromium plus vision models to capture both DOM and visual style.