Open-source component library for building AI-native applications. Beautiful, composable primitives for chat, streaming, and multimodal experiences.

Wispr Flow: Dictation That Works Everywhere — Stop typing. Start speaking. 4x faster.

Stop typing. Start speaking. 4x faster.

Hey everyone — thanks for being here. Most component libraries stop at general-purpose primitives. That’s fine for a lot of apps, but for AI products, you still end up composing the whole experience yourself. I wanted a library that meets those patterns head-on, so you’re not rebuilding the same components from scratch every time. Nexus UI is what I wished existed: beautiful, accessible, copy-paste components built for AI apps — streaming, multimodal inputs, and patterns that play nicely with stacks like the Vercel AI SDK (and room to grow for voice and more). Early on, it was “just a few primitives.” As it grew, the focus shifted to distribution and quality: a proper registry, docs you can actually follow, and shipping components only when they felt production-ready. If you’re building an AI product and want components that match how these products really work, I’d love your feedback, especially what’s missing for your stack.

@victorwill__ This is a smart focus — AI apps do have pretty distinct composition patterns that general component libraries miss. The emphasis on shipping only when production-ready feels like the right call too, especially for something developers are copy-pasting into production. Curious if you've found certain AI patterns (like prompt management or token counting UX) harder to standardize than others.

Wispr Flow: Dictation That Works Everywhere — Stop typing. Start speaking. 4x faster.

Stop typing. Start speaking. 4x faster.

Nexus UI

Beautiful, composable components for building AI apps

Beautiful, composable components for building AI apps