Community

The Neutro Ecosystem

Package	Description	Status
@neutro/view	Fine-grained reactive view engine	v0.1.0
@neutro/form	Zero-dependency reactive form engine	v0.3.0
@neutro/fluid	Physics-grounded glass material system	Coming soon

Filing Issues

• Found a bug? Open an issue with a minimal reproduction
• Have a feature request? Open an issue with the use case and expected behaviour
• Question or discussion? Start a discussion

Support the Project

If @neutro/view saves you time, consider buying me a coffee — it keeps the packages maintained and the documentation up to date.

---

FAQ

Core Concepts

Why signals instead of a virtual DOM?

Signals record which DOM nodes read which values at mount time. When a signal changes, only those specific nodes update — no diffing pass, no component re-render, no scheduler. The update cost is proportional to what changed, not to the size of the component tree.

What's the difference between .nv files and the tagged template?

.nv files use compiler erasure: you write bare signal reads (count) and assignment-form writes (count = count + 1), and the compiler rewrites them to accessor calls before execution. The tagged template has no compiler step — every reactive value must be an explicit thunk (() => count()) and every write must call .set(). Both compile to the same TemplateIR at runtime.

Do I need a build step?

Only for .nv files, which require esbuild + nvPlugin. The tagged-template path (createHtmlTag + mount) works in plain TypeScript with no build plugin — serve with Vite or any TypeScript-capable dev server.

Can I use @neutro/view with React, Vue, or Svelte?

The reactive core (@neutro/view/core) is framework-agnostic and has zero DOM dependency. You can use signals anywhere. Mounting nv components alongside other frameworks requires a shared DOM target — it works but there are no official adapters in v0.1.0.

Reactivity

Why does my effect not run immediately after a signal write?

Effects are microtask-scheduled. They run after the current synchronous code finishes. Use flushSync() if you need the DOM to reflect new signal values before the next line.

What happens when I write to multiple signals in a row?

Each write schedules effects independently by default. Use batch(() => { ... }) to defer all scheduling until the batch returns — dependents see only the final state and effects run once.

Can I read a signal without creating a dependency?

Yes — wrap the read in untrack(() => signal()). Reads inside untrack are not registered as dependencies of the enclosing computed or effect.

TypeScript

Is TypeScript supported?

Yes. All exports from @neutro/view/core and @neutro/view/renderer are fully typed. The .nv compiler output requires // @ts-nocheck in importer files at v0.1.0 — this is planned for resolution in v0.5.0.

Why do I need // @ts-nocheck when importing .nv files?

The nv compiler emits JavaScript modules, not TypeScript declarations. Until the compiler generates .d.ts files (planned for v0.5.0), TypeScript cannot resolve the types of .nv exports, so you must suppress the checker in the importing file.

Runtime

Can I use @neutro/view/core in Node.js?

Yes. The core has zero DOM dependency and runs in any JavaScript environment — Node.js, Web Workers, Deno, Bun. The renderer (@neutro/view/renderer) requires a DOM environment (browser or jsdom).

How do I test components?

Write reactive logic tests against @neutro/view/core in plain Node.js with vitest — no jsdom needed. For renderer tests that touch the DOM, use vitest + jsdom. Call flushSync() after signal writes to synchronously flush effects before asserting DOM state.

What's planned for v0.5.0?

Async primitives, context API, a module-level store, SSR support, TypeScript declarations for .nv outputs, and the path-scoped sidebar / Guides section on this docs site.