Harsh Patel

GenUI. Builder. Building.

[Harsh Patel @harshpatel071]

I’ve been working on the problem of making generative UI work at scale, and the deeper I’ve gone, the more I’ve found myself questioning what “generative UI” actually means. Here are my thoughts... https://t.co/bR1Zt4XIZj

[Harsh Patel @harshpatel071]

@elijahmuraoka_ @tomoji_ai Need to start 😭 congrats 🎉

[Corey J. Gallon @CoreyGallon]

@forwardsecrecy If allowing your agents access all of the sites that you do, on your behalf, is important to you, I spoke about how to do just that at AI Engineer World’s Fair last month.

They’ve published my talk, “The Dark Arts of Web Automation” here:

https://t.co/ikiOHTGsnB

[Harsh Patel @harshpatel071]

@deployengineer @mastra @calcsam Agreed, both books are great reads

[Harsh Patel @harshpatel071]

What problems are people facing with generative user interfaces, or dynamic interfaces that they use on prod?

[Harsh Patel @harshpatel071]

@LeoOliemans91 Agreed

[Harsh Patel @harshpatel071]

Anthropic is on board too.

AG-UI now includes an integration for Claude Managed Agents, another sign that the agent-to-frontend layer is becoming a serious part of the AI stack. https://t.co/uvF10tZS4P

[Harsh Patel @harshpatel071]

I’ve been working on the problem of making generative UI work at scale, and the deeper I’ve gone, the more I’ve found myself questioning what “generative UI” actually means. Here are my thoughts... https://t.co/bR1Zt4XIZj

[Harsh Patel @harshpatel071]

@elijahmuraoka_ Congrats!

[Bhaulik Patel @deployengineer]

https://t.co/AJAiu6tmTw

[Bhaulik Patel @deployengineer]

Short self-serve crash course on Hermes Agent to build your understanding. Linking in comments:

[Harsh Patel @harshpatel071]

Claude Mythos benchmarks blow my mind.

[Harsh Patel @harshpatel071]

I’m open-sourcing rxjs-ai.

I built an internal abstraction to handle LLM streaming (cancel, retry, throttle, race) using RxJS, since that’s already what I use for streams elsewhere. It’s been handy in my project, so I’m open-sourcing it as rxjs-ai for anyone who wants the same building blocks.
Try it out: https://t.co/qypset9pJa

GitHub: https://t.co/JFBF9OBPU2

Docs: https://t.co/vmSceAEI8L

[Harsh Patel @harshpatel071]

@angular https://t.co/3B6uDQMER4 😉

[Harsh Patel @harshpatel071]

@denicmarko https://t.co/xVXlCxUw3I

[Harsh Patel @harshpatel071]

The "Angular Renaissance" is real.
A while ago, I released the first version of AI-NG. To my absolute shock, it caught the eye of Minko Gechev (Google/Angular Core Team), who reposted it. It was an experimental tool designed to bridge the gap between AI and production-grade Angular code.

That one moment of validation from the lead of Angular changed everything.

In the world of building, there’s often a "trough of sorrow" where you wonder if what you’re creating actually matters. That traction didn't just give me views; it gave me the fuel to keep going when I was ready to pivot.

Today, I’m thrilled to announce that AI-NG (https://t.co/JS8y0UfS0G) v2 is officially here.
We are living through what @mgechev explains about the Angular Renaissance (https://t.co/UeKm9p8Wf7) - a complete reimagining of the framework with Signals, Standalone Components, and Deferrable Views. AI-NG v2 is built specifically for this new era. It’s no longer just a "no-code" experiment; it’s a sophisticated engine that generates modern, high-performance Angular code following the latest best practices (including the Minimal Component Pattern). It also has a learn mode for devs to catch up with the Angular changes.

What’s new in v2?
Signals-First Architecture: Everything generated is optimized for Angular’s new reactive primitive.
Minimalist UI/UX: A modern DX, the tool itself feels as fast as the code it produces.
Learn Mode: Beyond just generating code, it helps developers understand why certain patterns are used in the modern Angular ecosystem.

Building in public is a rollercoaster, but the community support has been the ultimate stabilizer. If you're riding the Angular Renaissance wave, I'd love for you to check out what I've been working on at https://t.co/JS8y0UfS0G.

Let’s build the future of the web, one Signal at a time. ⚡️

@GoogleAI @Google
#Angular #WebDevelopment #AI #SoftwareEngineering #AngularRenaissance #BuildInPublic

[Harsh Patel @harshpatel071]

AI-NG V2 is launching
Think lovable for Angular, plus a build-in learn mode!

AI-NG generates production-ready, clean Angular code.
Learn and build with the Latest Angular versions! try it here: https://t.co/JS8y0UfS0G

Deeply grateful to the @angular team for making Angular more AI-ready than ever. The team’s work on AI compatibility is commendable.

https://t.co/iSetsrYa8J… via @ProductHunt - show some love

Is it upwards for Angular from here? You bet. 📈
#BuildInPublic

[Harsh Patel @harshpatel071]

AI-NG V2 is launching
Think lovable for Angular, plus a build-in learn mode!

AI-NG generates production-ready, clean Angular code.
Learn and build with the Latest Angular versions! try it here: https://t.co/JS8y0UfS0G

Deeply grateful to the @angular team for making Angular more AI-ready than ever. The team’s work on AI compatibility is commendable.

https://t.co/iSetsrYa8J… via @ProductHunt - show some love

Is it upwards for Angular from here? You bet. 📈
#BuildInPublic

[Harsh Patel @harshpatel071]

AI-NG V2 is launching
Think lovable for Angular, plus a build-in learn mode!

AI-NG generates production-ready, clean Angular code.
Learn and build with the Latest Angular versions! try it here: https://t.co/JS8y0Ufkb8

Deeply grateful to the @angular team for making Angular more AI-ready than ever. The team’s work on AI compatibility is commendable.

https://t.co/TtmFzO6IrT via @ProductHunt - show some love

Is it upwards for Angular from here? You bet. 📈
#BuildInPublic

@mgechev @pkozlowski_os @synalx @jelbourn @Jean__Meche