hey everyone,
We’re RightNow AI team and we’re Nvidia inception member https://www.rightnowai.co/
we’ve been working on the first cuda ai code editor to make everyone able to write reliable fast cuda kernels using AI
Our editor handel:
- cycle-accurate gpu emulator (98% accuracy) so you can test multi-gpu without the hardware
- real-time profiling and benchmarking while you code
- ai that generates optimized kernels based on actual microarchitecture understanding
- integrated debugging and performance analysis
runs on local LLMs so your code never leaves your machine and it’s completely free
built on vscode fork so it should feel familiar
still early and needs a lot of work. would love to hear what features would actually be useful for you or what’s missing
