Blog
Design extraction, explained with working code
Tutorials and engineering notes on pulling design tokens, brand data, fonts, SVGs, and clean code out of live websites. DIY approaches included - the API is the shortcut, not the prerequisite.
How to Connect the MiroMiro MCP to Claude, Cursor, ChatGPT and Codex
Step-by-step connector setup for every client: paste one URL into Claude web, one-click install in Cursor, developer mode in ChatGPT, one command in Claude Code. Works without a key in demo mode.
Read the postAdd Design Extraction to Claude Code with MCP
Install the MiroMiro MCP server in two commands and Claude Code can pull any website's tokens, brand, fonts, and section code instead of guessing.
Read the postBest Brand Data APIs in 2026 (Logos, Colors, Fonts)
Brandfetch, Logo.dev, Clearbit, and MiroMiro compared honestly on depth and pricing - when a logo API is enough and when you need the full design system.
Read the postBest Design-to-Code APIs and Tools in 2026
The design-to-code landscape in 2026 sorted by input - Figma files, screenshots, prompts, or live URLs - and which tool honestly wins for each job.
Read the postBest Design Token Extraction Tools in 2026
Six real ways to pull design tokens from a live website in 2026 - free analyzers, a DIY script, a browser extension, and an API - honestly ranked.
Read the post5 Best Firecrawl Alternatives in 2026
Five real Firecrawl alternatives in 2026, matched to the job: cheap markdown, open source, site automation, agent search, or design context for AI.
Read the postDesign Token Types: Every Token a Website Exposes (Reference)
All 17 design token types a website exposes - the CSS properties each comes from, what good output looks like, and the gotcha that silently corrupts it.
Read the postHow to Detect and Extract the Fonts a Website Uses
Three ways to detect which fonts a website really uses - computed styles, document.fonts, and network sniffing - and why font-family lies to you.
Read the postHow to Extract Brand Assets from 1,000 URLs in Bulk (Python)
Bulk-extract brand data and design tokens from thousands of domains in Python - async code with rate limiting and retries that survives a real URL list.
Read the postHow to Extract Design Tokens from Any Website in Python
Extract colors, fonts, spacing, and shadows from any website with Playwright in Python - working code, the gotchas that bite, and the one-API-call shortcut.
Read the postHow to Extract a Website's Color Palette Programmatically
Two working ways to get a site's color palette in code - screenshot quantization vs reading the CSS - and which one returns the actual brand colors.
Read the postEverything in these posts, as one API call
Design tokens, brand data, fonts, SVGs, images, and section code from any URL. 100 free credits a month, no card.