v0 by Vercel
Vercel's generative UI tool — turn a prompt, screenshot, or design into production React, Next.js, and Tailwind code
AI design tool that turns prompts into interactive prototypes built from your team's real design system
Magic Patterns is an AI design tool for product teams that generates interactive prototypes from your real design system rather than generic UI. The free tier includes 100 monthly credits; Starter is $20 per seat per month for 1,000 credits and Business is $100 per seat per month for 5,000 credits, with a single generation averaging about 25 credits. Best for design and product teams that need Figma and GitHub round-trips — not a full app builder.
Magic Patterns is a web-based AI design tool aimed at product teams rather than solo builders. You describe a feature — or paste a user story or screenshot — and it generates an interactive, clickable prototype instead of a static mockup. The differentiator is its Design Systems feature: before generating anything, you seed a system from your existing Figma library, a GitHub repo, a published NPM component package, or your live website. Generations then reuse those real components, color tokens, typography, and styling rules, so prototypes look like your product instead of generic AI output. Built-in systems such as Base, Shadcn, and MUI are available if you have nothing to import.
The second half of the product is handoff. Finished designs export to Figma through a plugin, sync two-way with a GitHub repository, or download as a code zip. A Magic Patterns MCP server connects designs directly into AI coding tools, and a v3 REST API lets you create designs, send prompts, read and write artifact files, and publish changes programmatically. Billing runs on credits that scale with request complexity: a single generation averages about 25 credits, a small prototype of six to ten iterations lands near 180, and unused monthly credits roll over for one month. Paid plans can enable on-demand usage at $0.02 per extra credit once the monthly allowance is gone. The company, operating as North Park Labs, was founded in 2023, went through Y Combinator, and announced a $6M Series A led by Standard Capital in November 2025.
Magic Patterns fits teams whose bottleneck is alignment before engineering starts — where the goal is a prototype stakeholders and customers can click through, built in the company’s own design language, not a deployable app. The free tier’s 100 monthly credits are enough to evaluate the generation quality but not to do real work, since watermarks stay on and every export path is gated behind a paid plan. It is a poor fit if you want an agent to build and deploy a production application; tools like Lovable and Bolt target that job, and for the actual production code work that follows a prototype, an agentic coding tool such as Claude ranks above the design-side output here.
Starting price: $20/mo · Free tier: yes · Model: freemium
Price history tracked from June 2026
| Plan | Price | Includes |
|---|---|---|
| Free | $0 | 100 monthly credits · AI generation in the web editor · Designs carry a watermark · No on-demand usage — generation stops until the monthly refresh |
| Starter | $20/seat/mo | 1,000 monthly credits · $17/seat/mo on annual billing (15% off) · On-demand usage at $0.02 per extra credit · Design systems, GitHub sync, and Magic Patterns MCP · No watermark, centralized billing · Team workspace up to 10 paid seats |
| Business | $100/seat/mo | 5,000 monthly credits · $85/seat/mo on annual billing (15% off) · Team workspace for 11+ paid seats · Faster AI models · Self-serve SSO, user roles and permissions · Usage reporting |
| Enterprise | Custom | Custom credit pools with shared credit pooling · SCIM and audit logs · Technical account manager and priority support · Custom contracts and flexible invoicing · Annual commitment required |
| Pros | Cons |
|---|---|
| Design-system import from Figma, GitHub, NPM, or a live website means output resembles the existing product instead of generic AI styling | Hard credit ceilings — Starter's 1,000 monthly credits cover roughly five to six small prototypes at about 180 credits each, after which on-demand billing at $0.02 per credit applies |
| Multiple genuine handoff paths — Figma export, two-way GitHub sync, code download, MCP, and a REST API | Output is prototype-grade: the generated React code is a starting point an engineer adapts, not a production application you ship as-is |
| Unused monthly credits roll over to the following month rather than being forfeited at renewal | Per-seat pricing scales awkwardly for larger teams, and Business at $100/seat/mo is a 5x jump from Starter for the same core editor |
| Credit consumption is published with averages per iteration count, so session cost is estimable before starting | The free tier watermarks designs and excludes every export path — Figma, GitHub sync, and MCP all require a paid plan |
| Overlaps heavily with cheaper or free prompt-to-UI rivals if you do not need the design-system fidelity |
Vercel's generative UI tool — turn a prompt, screenshot, or design into production React, Next.js, and Tailwind code
AI-powered UI design tool that generates multi-screen prototypes from text prompts
AI-powered UI/UX design tool that turns text prompts, screenshots, and sketches into editable wireframes and mockups for non-designers
AI app builder that turns text prompts into deployable web apps and SaaS products
AI-powered web development platform that builds full-stack apps from a chat prompt
Google Labs' free AI design canvas that turns text prompts into UI screens, Figma exports, and front-end code
See all Magic Patterns alternatives →
Yes, there is a free tier with 100 monthly credits. Free designs carry a watermark, and on-demand usage is not available, so generation stops once credits run out until the next monthly refresh. Paid plans begin with Starter at $20 per seat per month, which includes 1,000 monthly credits and unlocks the export options.
Starter is $20 per seat per month with 1,000 monthly credits, and Business is $100 per seat per month with 5,000 credits. Annual billing takes 15 percent off, bringing those to $17 and $85 per seat per month respectively. Enterprise is custom priced and requires an annual commitment.
Credits are consumed by AI generations and scale with request complexity, so a color tweak costs less than a full multi-page prototype. A single generation averages roughly 25 credits, a six-to-ten iteration prototype lands near 180 credits, and a project of 21 to 50 iterations averages about 600 credits. Manual edits, exports, sharing, and Fix with AI consume nothing.
Yes. Unused monthly credits roll over to the following month and then expire after one month. If a plan grants 100 credits and 50 remain at the next renewal, the balance becomes 150 credits, and those carried-over 50 expire if they are still unused by the renewal after that.
Yes. Designs export to Figma through a plugin, and GitHub sync offers two-way push and pull with a repository. You can also download the code as a zip file, connect the MCP server to Cursor or Claude Code, or copy a design's code as a prompt for another AI tool. These paths require a paid plan.
Yes. The v3 REST API creates designs, sends prompts, reads and writes artifact files, and publishes changes programmatically. Requests authenticate with an API key in the x-mp-api-key header, and the same key also works for the MCP server. A legacy v2 endpoint still exists but is being deprecated.
No. It targets prototyping and design handoff rather than shipping production applications. The generated code is a starting point an engineer adapts, which is why the product ships an MCP server and an integration skill for AI coding tools. For production coding work, a dedicated agentic tool such as Claude is the stronger choice.