RightNow AI
#1 GPU AI Code Editor
Start new thread
Jaber Jaber

RightNow - AI code editor for GPU kernel development

by
RightNow AI
•
RightNow AI is the first GPU-native code editor supporting CUDA, Triton, CUTE, and Tilelang. Features agentic hardware-aware AI (Forge), cycle-accurate GPU emulator with 98% accuracy across 86+ NVIDIA architectures, real-time profiling with Nsight Compute integration, and line-by-line performance analysis. Write, optimize, and profile GPU kernels in any DSL, all in one editor. Supports all NVIDIA GPUs Free to download with pro features available.

Add a comment

Replies

Best
Jaber Jaber
Maker
📌
GPU development shouldn't require switching between five tools just to understand why your kernel is slow RightNow AI is the first GPU-native code editor. It brings real-time profiling, a cycle-accurate emulator, and AI that actually understands GPU kernels—all in one place. profile without leaving your code. emulate H100s without the hardware. ask optimization questions in plain english We support CUDA, CUTE, Triton, and Tilelang free to download. give it a try, and join our discord if you have feature requests: https://discord.com/invite/sSJqg...
Masum Parvej
💎 Pixel perfection

@jaber23 Cycle accurate emulation feels powerful. How do you keep performance overhead low? Maybe lightweight modes could help.

Jaber Jaber
Maker

@masump we're ptx-based, ptx simulation is inherently fast while still giving you cycle-accurate insights across architectures

ISTIAK AHMAD
Been waiting for something like this! cuda dev has been such a pain with all the context switching between tools
Jaber Jaber
Maker

@istiakahmad that's why we build it:D

Malith Gamage

This look great!

Jaber Jaber
Maker

@malithmcrdev thanks!!

Abdul Rehman

Does it help identify memory bottlenecks automatically, or do I still need to hunt them down?

Alexander Gifford

Being able to leverage local LLMs for coding more efficiently is awesome.

Tomo(Patch)

Huge congrats on shipping this.

“GPU-native” editor + real-time profiling + H100 emulation in one place feels like a massive DX upgrade vs juggling Nsight + scripts + docs.

Really curious to see how close the emulator gets to real hardware in practice.

Jaber Jaber
Maker

@patch_app Thanks Tomo<3

Aditya Raj

GPU emulation is notoriously slow. How close is your emulator’s behavior to real hardware, especially for warp scheduling and memory access patterns?

Raju Singh

This is solving such a critical pain point in GPU development. The 98% accuracy emulator + real-time profiling combo is genuinely powerful.

Question: What's your user acquisition strategy? Are you targeting individual CUDA devs or enterprises with large-scale GPU infrastructure? Also curious about pricing model - per license or cloud-based consumption?

Love the focus on developer UX for a niche but growing market. Great execution! 💪