mmo-chip: Multiplayer CMOS Standard Cell Chips Reverse Engineering Tool

A new open-source tool called mmo-chip enables reverse engineering of CMOS standard cell chips from die photographs, allowing users to import gigapixel images, annotate vias and traces with optional machine learning assistance, and extract circuit schematics and Verilog netlists. The tool, still in early development, includes a die viewer, cell reverse engineering features, and a Python-based ML sidecar for via/trace detection.

A tool for reverse-engineering integrated circuits from die photographs, with a special focus on CMOS Gate Array/Standard Cell chips. Import and tile gigapixel die shots, annotate them vias, traces, standard cells with optional ML assistance, and extract the circuit, per-cell transistor and logic-gate schematics, and a die-level Verilog netlist. Die viewer — navigate the whole die, with placed cells, nets, and I/O pins. Cell RE — annotate a cell's layers; transistors, gates, and logic are inferred. Die shots in those screenshots are from InfoSecDJ http://infosecdj.net/ . frontend/ Vite + React + TypeScript UI die viewer, cell RE, schematics, Verilog backend/ Node + TypeScript API — image import, tiling, JSON persistence, WebSocket shared/ Shared TypeScript types the annotation schema ml/ Python U-Net for via/trace detection + FastAPI prediction sidecar data/ Imported dies, tile pyramids, and ML exports Node ≥ 20 npm workspaces — required. Python ≥ 3.10 — only for the ML sidecar assisted via/trace annotation . JS/TS workspaces frontend + backend + shared npm install ML sidecar optional python3 -m venv ml/.venv ml/.venv/bin/pip install -r ml/requirements.txt npm run dev Starts the backend http://localhost:3001 http://localhost:3001 , the frontend http://localhost:5173 http://localhost:5173 , and the ML sidecar together. Open the frontend URL. Without the ML sidecar, run just the two TypeScript apps: npm run dev -w backend npm run dev -w frontend Build and test: npm run build npm test IMPORTANT NOTE: This software is still very early in development and has been tested only locally firewalled . Use at your own risk.