Extract Any Brand DNA Instantly

StyleExtract is an AI-native brand extraction platform that analyzes any public website URL and produces a complete, structured brand infrastructure payload. The Full Brand Extraction includes DESIGN.md (AI-ready style tokens), AGENTS.md (explicit AI coder guide), .cursorrules (Cursor IDE workspace rules), .mdc files (Cursor native rules with glob patterns), llms.txt (high-density LLM context), a cleaned HTML scaffold, CSS custom properties, Tailwind v3/v4 configs, Shadcn UI theme, Figma tokens in W3C DTCG format, a branded DOCX template, a branded PPTX presentation, and a 17-page PDF brand book.

DESIGN.md is a structured Markdown file with YAML front-matter containing exact design tokens — colors in HEX/RGB/CMYK/HSL, typography scale, spacing grid, border-radius tokens, shadow levels, and component patterns. It is optimised for use with AI coding agents like Claude, Cursor, GitHub Copilot, and GPT-4. When you add DESIGN.md to your project root, any AI agent that reads Markdown context will generate UI components that exactly match the target brand's design system.

AGENTS.md is an explicit documentation file formatted specifically for external AI coders. Unlike DESIGN.md which is token-focused, AGENTS.md defines behavioral contracts: what components MUST look like, what the agent MUST NOT do, integration checklists, and machine-readable JSON color palette blocks. It works with any AI agent that accepts system context — Claude, GPT-4, Gemini, or custom LLM pipelines.

.cursorrules is a workspace context file pre-configured with your brand colors, typography, spacing scale, and component rules so that Cursor IDE always generates on-brand UI code. .mdc files are Cursor's native rule format with YAML frontmatter — they support glob patterns so you can scope brand rules to specific file types (e.g., only *.tsx files). Both are included in the Full Brand Extraction and require no setup beyond dropping them in your project.

llms.txt is a stripped, high-density markdown file that compresses the brand's visual identity into the smallest possible context footprint. Unlike DESIGN.md which is comprehensive, llms.txt is optimised for paste-in use with token-limited LLM prompts — each section is terse and structured for maximum information density per token.

The Developer API is a subscription-gated REST API ($59.99/month or $600/year) that lets corporate users and AI agents interface directly with StyleExtract. Authenticated with bearer API keys (sk_style_...), it provides three endpoints: POST /api/v1/extract (trigger a new extraction), GET /api/v1/extractions (list all past extractions), and GET /api/v1/extractions/{id} (retrieve individual files like DESIGN.md or .cursorrules without burning your daily quota). Rate limited to 1 new extraction per calendar day per API key.

You get 15 files: DESIGN.md, SPEC.md, llms.txt, HTML scaffold, AGENTS.md, .cursorrules, brand.mdc, CSS Custom Properties (.css), Tailwind Config v3 (.js), Tailwind v4 @theme (.css), Shadcn UI theme (.css), Figma Tokens Studio JSON, W3C DTCG design tokens JSON, branded Word DOCX template, branded PowerPoint PPTX presentation, and a 17-page PDF brand book. All files download natively with correct extensions and MIME types.

StyleExtract uses a headless Chromium browser to visit your target URL, compute CSS styles across every visible element, and rank colors by brand relevance. Colors are deduplicated using Euclidean RGB distance so you get the true distinct palette. Saturated, intentional brand colors score higher than neutral backgrounds. The extraction also captures typography, spacing tokens, border-radius patterns, shadow values, logo URL, and semantic page structure.

No account is required for one-time extractions. Paste any URL, run a free extraction, and pay once to unlock the Full Brand Extraction — no login, no subscription. The Developer API ($59.99/month) requires an account for API key management and extraction history. B2B subscribers can log in at /dashboard to view their full extraction history and manage API keys.

Yes. The Figma Tokens Studio JSON export is structured for direct import into the Tokens Studio plugin for Figma. It includes color tokens, font families, a full type scale from display to xs, font weights, line heights, spacing tokens on an 8px grid, border radius tokens, and box shadow elevation tokens. The W3C DTCG format is also included for use with Style Dictionary or any standards-compliant token pipeline.

StyleExtract works on any publicly accessible website. It renders each page using a real Chromium browser, so it fully supports JavaScript-heavy SPAs, React apps, Vue apps, Next.js sites, and standard HTML. Sites using aggressive bot-blocking such as Cloudflare CAPTCHA challenges may not be accessible.

Yes. StyleExtract uses docx and pptxgenjs to compile scraped brand data into valid binary XML-compressed Office files — not text files with a renamed extension. They open correctly in Microsoft Word, PowerPoint, Google Docs, and Keynote without format errors.