Open-source tools for local AI
Run AI on your own machine
The same local-AI toolchain that ships in Shadowfetch Linux installs on any Debian Testing box. Local language models and a private chatbot run on your own hardware -- no cloud account, no API keys, and no need to switch your OS.
Top selling apps
All apps →
One command, any Debian
The toolchain lives in shadowfetch-defaults. Add the signed repository and install it, and the assistant, model and GPU helpers are set up for you -- without reinstalling your system.
sudo apt install shadowfetch-defaultsThe toolchain
Each is one command. Start with the private chat, or inspect your hardware and choose a model first.
shadowfetch-aiA private local chatbot using Ollama with an Open-WebUI front end, running entirely on your hardware.
shadowfetch-llmPick and compare local models matched to your GPU and memory, so you run the largest model your machine can actually hold.
shadowfetch-gpuInspect graphics hardware and available acceleration before changing drivers or selecting a model.
shadowfetch-healthRun plain-language system checks before and after installing local model software.
What runs on your hardware
We benchmark local models on our own card — a single 16 GB GPU — and publish the numbers: tokens per second, memory used, and whether a model fits entirely on the GPU. See what your hardware will actually do before you install anything.
It stays on your machine
The toolchain runs entirely on your hardware via Ollama. Ollama listens only on localhost and the firewall blocks port 11434 -- your models never leave the machine. There is no account, no telemetry, and nothing to sign in to.
Before you install
Three questions that come up before the first command, answered from what the packages actually ship.
Do I need an NVIDIA GPU, or does AMD / CPU-only work?
No NVIDIA card is required. The proprietary NVIDIA bundle (shadowfetch-nvidia) is optional, and a first-boot hook removes it when no NVIDIA GPU is detected, so it is safe on non-NVIDIA hardware. Firewatch treats NVIDIA telemetry as optional (a Debian Recommends, never a hard dependency). The hardware scanner reads any GPU it finds through Vulkan as well as nvidia-smi, and the local runtimes run on the CPU when there is no supported GPU, so you simply run smaller models.
What does shadowfetch-defaults install?
It applies the Shadowfetch defaults to new user accounts through /etc/skel (kdeglobals, kwinrc, a Konsole profile and a fastfetch config), adds a profile.d snippet that prints a fastfetch banner on your first interactive shell, and ships the Shadowfetch Control Center, safe updater, and health diagnostics. It depends on fastfetch, python3 and python3-pyqt6.
Which Debian does this target?
Debian Testing, on amd64. Shadowfetch is a Debian-testing derivative: Debian provides the base and Shadowfetch adds the curated layer on top. The signed APT repository ships a single main component built for amd64, so the toolchain installs on any current Debian Testing box without switching your operating system.