LLMs are autocomplete for intent.
As a UI engineer, that changes everything.
Anywhere you can predict or complete what a user is doing, AI belongs there.

👨‍🏫 We are launching workshops for React and Angular. I will personally teach you everything I know about using LLMs to build generative user interfaces. They are hosted remotely, with the first dates set for October. Best part? The revenue supports the open source project!

💻 Huge shoutout to the team on this, especially @blove.dev. We've launched a complete redesign of the Hashbrown docs site, complete with tons of new documentation. There's still plenty to write, but we're making big strides in our docs game with this release.

📝 Our new prompt template literal function allows you to write UI examples in your system instructions. Hashbrown validates them against your component schema, then downlevels the example into the underlying JSON representation. Result? Much better prompts!

🦙 You can now use open-weight models via our Ollama adapter. Setting it up in your backend is just as smple as our other adapters, and we support Ollama Turbo to run models on their infrastructure.

We got Hashbrown working against gpt-oss:120b. Generative user interfaces powered by 100% open source code. No API keys, subscriptions, or external service providers in the loop.

Good morning to the Angular community! 👋🥔

Post a perfect album from the 90s that isn’t Nirvana, Pearl Jam, Soundgarden, or Alice In Chains.

It’s not the prettiest line chart, but it is…

- Built with Angular components and directives
- Fully accessible
- Rendered to SVG
- Has zero dependencies
- Leverages incremental hydration to defer loading interactivity (i.e. tooltips)
- Styled with CSS variables for easy customization