Tools
Small, focused extensions for people running models on their own hardware. No accounts, no telemetry, no cloud round-trip.
v0.2.3 · Windows · Linux · macOS
VRAM, temperature and utilization in your status bar.
Real-time GPU monitoring without leaving the editor. Click through to a detail panel for memory, temperature, and utilization. Supports NVIDIA, AMD, Intel, and Apple Silicon across Windows, Linux, and macOS.
01code --install-extension SavicLabs.savic-labs-gpu
v0.10.9 · llama.cpp · on-device
Your local models, inside GitHub Copilot Chat.
Point Copilot Chat at llama.cpp running on your own machine. Zero configuration — models are discovered automatically. Supports thinking and reasoning output, tool calling, and a vision proxy.
01code --install-extension SavicLabs.savic-labs-local-ai
Why we publish these
These exist because we needed them. Running models locally means knowing what your hardware is doing and keeping inference off someone else's servers — so both tools do exactly one job and never phone home.
They are free, and they are the same instruments we use while building NikkiOS.