// the find
BuilderIO/mitosis
Write components once, run everywhere. Compiles to React, Vue, Qwik, Solid, Angular, Svelte, and more.
Mitosis lets you write UI components once in a JSX-like subset and compile them to React, Vue, Angular, Svelte, Solid, Qwik, and a handful of others. It's primarily built by Builder.io to power their visual editor across frameworks, but it's open enough to use for maintaining a multi-framework design system without duplicating source code.
The IR (intermediate representation) approach is sound — components are parsed into a framework-agnostic AST, and per-target generators handle the idiomatic output. This means generated React looks like React and generated Svelte looks like Svelte, not a lowest-common-denominator wrapper. The e2e test suite actually runs compiled output against real framework runtimes with Playwright, which is the right way to catch regressions in a compiler. The Figma integration is a genuine differentiator — going from design tokens to framework-native components is a real workflow, not a demo. Signals support (visible in the e2e components) means it tracks the modern reactivity direction across Solid and Qwik rather than being stuck on 2020-era patterns.
The constraint that your source must stay within Mitosis's supported JSX subset is tighter than it looks — complex hooks, render props, context providers, and anything idiomatic to a specific framework get flagged or silently dropped. The feature matrix doc (`docs/feature-matrix-draft.md`) is literally a draft, which tells you the coverage gaps are real and not well-documented. If you hit an edge case, you're reading compiler source to understand why output is wrong. The 'write once' promise breaks the moment frameworks diverge on something meaningful — Svelte 5 runes and Angular signals are structurally different enough that the generated output for stateful components will lag behind what you'd write by hand. There's also an implicit dependency on Builder.io's roadmap priorities; the Figma integration and visual editor use cases drive decisions more than pure library-author needs.