Skip to content
@Cirkitly

Cirkitly

AI-Accelerated Hardware Design Tools — From spec to silicon in seconds.

Cirkitly

AI-Accelerated Hardware Design Tools — From spec to silicon in seconds.

Cirkitly is a modular open-source stack of intelligent agents and tools that reimagines the way digital hardware is designed. We use AI to radically simplify chip and accelerator development — from natural language specs to synthesizable RTL, optimized CUDA/HLS kernels, and layout-ready netlists.

Whether you're a student, FPGA developer, or silicon architect — Cirkitly accelerates the hardware workflow with clarity, automation, and AI augmentation.


What Can Cirkitly Do?

Tool Description
spec2rtl Convert structured specs → RTL + testbench + sim script
rtl-copilot Prompt-based Verilog generator & smart FSM editor
picoplace Lightweight placer-router with AI hints and SVG layout visualizer
cuda-genie AI compiler that converts C/Python → optimized CUDA/HLS kernels
silicon-agent Integration orchestrator that runs spec → RTL → sim → layout → report

Why Cirkitly?

Hardware development is traditionally:

  • Slow
  • Manual
  • Fragmented
  • Opaque

Cirkitly fixes that by introducing intelligent automation and an end-to-end AI-first stack, giving you:

✅ Rapid prototyping
✅ Seamless integration
✅ Natural language interfaces
✅ Powerful abstractions


Tech Stack

  • Languages: Python, Verilog, CUDA, JavaScript
  • Tooling: Yosys, Verilator, Graphviz, LLVM, Magic VLSI
  • AI Models: GPT-4, custom prompt engineering, agent routing
  • Frontend: React (Monaco), Tailwind, FastAPI
  • Simulation: Verilator, GTKWave, cocotb (planned)

Projects Inside

Each repo is modular, open, and can be used independently or as part of the full stack.


Demo Projects

Project Spec Type Tools Used Output
UART Controller YAML spec2rtl, rtl-copilot Verilog + Testbench + VCD
Pipelined ALU Prompt rtl-copilot, picoplace SVG layout + sim
Matrix Mult Python cuda-genie, silicon-agent CUDA kernel + perf notes

Demos are coming soon to Personal YouTube channel


How to Get Started

git clone https://github.com/Cirkitly/spec2rtl.git
cd spec2rtl
pip install -r requirements.txt
python cli.py --generate examples/uart.yaml

Or try the full stack orchestration via silicon-agent (coming soon).


Contributing

We’re looking for contributors passionate about:

  • EDA, FPGAs, or compilers
  • LLMs and AI agents for engineering tools
  • Hardware-software co-design

Open issues, suggest tools, or contribute to one of the core repos!


Maintained by

  • Utkarsh Maurya (@pro-utkarshM)
  • The open-source hardware community 🚀

License

MIT for core agents. Some tools (e.g., yosys, verilator) follow their own licenses.


Website

Website: https://cirkitly.utkarshlabs.tech


"Hardware should be as easy to write and debug as software — Cirkitly makes that real."

Popular repositories Loading

  1. spec2rtl spec2rtl Public

    AI-Powered Hardware Design & Verification Copilot

    Python 19 2

  2. spec2test spec2test Public

    Cirkitly is your AI-powered QA copilot for hardware developers.

    Python 5 1

  3. code2test code2test Public

    Python 1

  4. .github .github Public

  5. PocketFlow PocketFlow Public

    Forked from The-Pocket/PocketFlow

    Pocket Flow: 100-line LLM framework. Let Agents build Agents!

    Python

  6. website website Public

    website for cirkitly

    HTML

Repositories

Showing 7 of 7 repositories

Top languages

Loading…

Most used topics

Loading…