Introducing Kombai, the first AI agent built for frontend development.
Kombai vastly outperforms SOTA models + generic agents in real-world benchmarks.
Watch it add a complex new feature to an OSS codebase with 500K+ lines of code.
More use cases and links in comments
We’re thrilled to welcome @Kombaico as a Diamond Sponsor of MUI! 💎
With their support, we’ll continue innovating, growing, and making better and faster UI tools for everyone.
So @Kombaico came out today with AI Agent for Frontend development, and ofcourse I had to use it to build a @reactjs demo. Let's put it to test.
Here's my honest opinion
Kilo code is gathering lots of buzz in the programming subreddits. So, I looked into their OpenRouter usage data vs. Cline to see who might be using them.
Interestingly, the model usage patterns on both tools are very different. Some interesting findings and questions:
1. Kilo code users are using Gemini 2.5 at nearly 4.5 times the rate of Cline users!
Last month, compared to Cline, Kilo code used ~1.7x tokens on Gemini 2.5 Pro but only ~0.4x on Sonnet 4.
Possible reasons for this:
->Kilo users like Gemini more because of cost, performance, or context length
->Because of Gemini's large context window, per user token consumption of Gemini is much higher
Kombai is the first coding agent for the front end.
I tried Kombai on Cursor, and I was very impressed with its understanding of context and the high quality of the code output.
Their benchmarks look incredible!
It works on :
- VS Code
- Cursor
- Windsurf
After months of building, learning, and refining with hundreds of amazing frontend devs, today feels special…✨
We’re launching the first-ever domain-specific agent for Frontend!
Built with care, passion, and your incredible feedback.
Can’t wait for you to try it! 🚀💙
Every ML team is about to 10x overnight.
The first agentic Machine Learning Engineer is here, and it will blow your mind!
Copilots help you write code, but that's about it.
People still have to do most of the work.
That's about to change!