Skip to content
aicoolies logo

Onlook Review: The Open-Source Cursor for Designers That Turns Visual Edits Into Real React Code

Onlook is an open-source visual editor for React and Next.js with 25.9K+ GitHub stars (Apache 2.0) that lets teams design directly in real codebases. The current site positions it as Cursor for Designers, with AI, visual editing, component/layer inspection, versioning, and hosted/team workflows. Pricing now supports self-hosted free open source, hosted Starter free with MCP-call limits, and custom Teams plans. Official current surfaces emphasize React, Next.js, Storybook, and shadcn/ui rather than broad Vue/Angular support.

reviewed by Raşit Akyol March 31, 2026

Documented evidence

rubric editorial-review-v1

This review is grounded in documented sources and repository analysis. It does not claim a unique hands-on reproducibility record.

Sources checked

Verdict

Onlook is a strong design-development collaboration tool for React and Next.js teams, especially when designers need to work against the real product instead of a separate mockup file. Its open-source repo has 25.9K+ stars and the official pricing page supports free self-hosting plus hosted Starter and custom Teams options. The main caution is framework scope: current official surfaces emphasize React, Next.js, Storybook, and shadcn/ui, so Vue/Angular claims should not drive adoption without fresh primary-source confirmation.

80/100

overall

Speed84
Privacy90
Dev Experience82

What Onlook Does

Onlook is an open-source visual editor for React and Next.js applications that has been dubbed the Cursor for Designers. With 25.9K+ GitHub stars and Apache 2.0 licensing, it bridges the gap between design and development by providing a visual canvas that works with real code-backed interfaces — not static mockups or a proprietary format. The current site emphasizes AI, code, and design side-by-side, with component/layer inspection, versioning, templates, and team workflows built around the idea that the product codebase is the source of truth.

Architecture and Workflow

The technical architecture is elegantly simple. Onlook instruments your served bundle with data-oid attributes at build time. When you visually edit an element, it uses the oid to locate the corresponding JSX, patches it, and triggers a hot module reload. The AI layer uses Morph Fast-Apply and OpenRouter to stream code edits via diff-match-patch, showing updates simultaneously in both the canvas and the file tree. The entire stack runs on CodeSandbox SDK with Bun runtime for fast installs and hot reloads. A local Git branch tracks changes under the hood with checkpoint rollback support.

The workflow collapses the traditional design-development gap. You connect a code-backed React or Next.js project, design on an infinite canvas with real components, and keep product UI closer to the implementation. Click elements to inspect structure, navigate layers, preview code side-by-side with the visual design, and iterate with AI assistance. The current official site lists React, Next.js, Storybook, and shadcn/ui as core compatibility surfaces, so teams should validate any broader framework expectations before adoption.

AI Integration and Collaboration

AI integration is conversational and context-aware. Describe changes in plain English — make this hero section darker on hover — and Onlook generates the React and Tailwind diff, applies it, and shows the result instantly. Drop images, mockups, or documentation into the conversation as context for better AI results. The AI can scaffold entirely new features: ask it to add auth with Supabase and it generates pages, hooks, and schema wired to your existing UI. This is not generic code generation — it understands your project structure and component patterns.

Collaboration features position Onlook for team adoption. Share your canvas and leave spatial comments — annotations that are positioned on specific elements, not floating in a separate thread. Work together on designs in real time. When ready, push changes directly to your repository, review diffs before committing, and your changes become a real pull request for engineers to review and merge. No handoff, no translation, no design specs that get misinterpreted.

Tailwind and Framework Support

The styling workflow is Tailwind-native. Visually edit and apply Tailwind classes with auto-completion and real-time previews. Build mobile-first with breakpoint previews and automatic media query generation. The editor toolbar lets you adjust Tailwind styles, directly manipulate objects, and experiment with layouts. All familiar hotkeys work — CMD/CTRL+Z and everything in between. Start from any Next.js and Tailwind template and the AI understands your patterns to generate matching components.

Current official framework positioning is focused on React and Next.js. The pricing and product pages also highlight Storybook and shadcn/ui, plus real design-system, theming, branding, branching, version-control, and component-tree workflows for teams. Because the current official surface does not clearly support broad Vue or Angular production use, buyers should treat Onlook primarily as a React/Next.js visual code editor unless they verify another framework path directly with the vendor.

Open Source and Limitations

The open-source nature is a genuine advantage. The stack is Apache 2.0 licensed, self-hostable from GitHub, and built transparently. Current pricing lists Starter as free for individuals and small teams with 200 MCP calls per day, unlimited projects, unlimited AI chat, and community support. Teams pricing is custom and adds templates, branching/version control, real design-system workflows, theming and branding, custom domains, advanced security, SSO, audit logs, admin controls, dedicated support, and custom integrations.

Limitations are practical. The tool is primarily optimized for React and Next.js product workflows, with official surfaces emphasizing Storybook and shadcn/ui compatibility. Visual editing of complex stateful components with hooks and context can be less intuitive than simple layout changes. AI code generation is not infallible, so complex features still require manual adjustment and code review. Hosted team features also use custom pricing, so organizations should expect a sales/demo step before standardizing around it.

The Bottom Line

Onlook represents a meaningful rethinking of how designers and developers collaborate on React applications. By working with code-backed interfaces rather than static mockups, it can reduce the handoff problem that slows product teams. For teams building React or Next.js applications, Onlook is a practical candidate for visual editing and AI-assisted UI iteration. Start with the free self-hosted or Starter path, verify the workflow against a real project, and evaluate Teams features when collaboration, security, and design-system governance become important.

Pros

  • Visual edits target real code-backed interfaces rather than throwaway mockups, reducing design-to-development handoff loss
  • 25.9K+ GitHub stars and Apache 2.0 open-source licensing provide transparency and a free self-hosted path
  • Official pricing lists hosted Starter free with daily MCP-call limits and custom Teams plans for collaboration, security, and support
  • AI chat, component/layer inspection, version history, branching, and real design-system features fit team workflows
  • Works well for React and Next.js teams using Storybook, shadcn/ui, and code as the source of truth
  • Teams can prototype and iterate visually while preserving code ownership and version-control review
  • Custom domains, theming/branding, SSO/security controls, audit logs, and admin controls are positioned for Teams plans

Cons

  • Current official surfaces emphasize React, Next.js, Storybook, and shadcn/ui; Vue/Angular support should not be assumed without fresh primary-source confirmation
  • Visual editing of complex stateful components with hooks, context, and conditional rendering can be less intuitive than layout changes
  • Hosted team features use custom pricing, so budgeting requires a sales/demo step rather than a public seat price
  • AI code generation for complex features may require manual adjustment and review before merging
  • Teams outside React/Next.js or without a code-backed design workflow may get less value than from traditional design tools

View Onlook on aicoolies

Pricing, platforms, and community stacks — explore the full tool page

Comparisons with Onlook

Onlook logo
Onlook
vs
Lovable logo
Lovable
vs
Tempo Labs logo
Tempo Labs

Onlook vs Lovable vs Tempo Labs — Visual React Builders & AI App Development Compared

Visual React development has become a distinct category in 2026, with tools that let designers and developers manipulate real React components through intuitive visual interfaces rather than writing code line by line. This comparison examines three platforms leading this shift: Onlook as an open-source visual editor that connects to your existing codebase, Lovable as a conversational full-stack app builder for rapid product creation, and Tempo Labs as a visual React IDE with multi-agent AI planning and a rich editing canvas.

Alternatives to Onlook

Cursor logo

Cursor

Top Pick

The AI-first code editor

AI-first code editor built as a VS Code fork that deeply integrates LLMs into every part of the development workflow. Features Tab autocomplete with multi-line predictions, Cmd+K inline editing, AI chat with full codebase awareness, and Agent mode for autonomous multi-file edits with terminal execution. Supports GPT-4, Claude, and more with automatic context from project files and docs. Includes privacy mode for SOC 2 compliance. The leading AI-native IDE with 100K+ paying users.

freemiumTelemetry

Self-hosted Heroku/Vercel alternative

Open-source, self-hostable PaaS alternative to Heroku, Vercel, and Netlify with 44K+ GitHub stars. Deploy static sites, APIs, full-stack apps, databases, and 280+ one-click services on your own VPS or bare metal via SSH. Features auto Let's Encrypt SSL, Git integration (GitHub/GitLab/Bitbucket/Gitea), S3 backups, Docker Swarm support, and a REST API for CI/CD automation. Self-hosted version is free forever with no features behind paywalls.

freemiumOpen Source

Convert any screenshot or mockup into clean frontend code with AI vision.

screenshot-to-code is an open-source tool that converts screenshots, mockups, and Figma designs into clean, functional HTML, Tailwind CSS, React, or Vue code using AI vision models. Powered by GPT-4V, Claude, and Gemini, it supports video-to-code and iterative editing workflows, making it a rapid scaffolding tool for frontend developers who want to skip the pixel-translation step entirely.

freemiumOpen Source

FAQ

How does Onlook visually edit React components without breaking AST structures?

Injects source-location identifiers into JSX and uses Babel/TypeScript AST traversal to update only targeted Tailwind classes while preserving code formatting and comments.

Does Onlook generate messy inline styles or clean Tailwind CSS utility classes?

Compiles visual modifications directly into native Tailwind CSS classes parsed from local tailwind.config.js, reflected via Hot Module Replacement (HMR) in sub-50ms.

How does Onlook's local desktop runtime compare to cloud builders?

Runs locally as an Electron app connected to your dev server (Next.js/Vite), writing standard React code to disk for collaboration via native Git PR workflows.

How does Onlook handle visual edits on reusable components and lists?

Distinguishes between call-sites and root definitions: single edits update passed className props, while component mode edits propagate globally without breaking map() bindings.

Sources & verification

Sources checked
Content verified

Verification dates are editorial checks. Routine CMS saves and automatic updatedAt timestamps do not advance them.