This project is based on large language models and a system-level automation framework. It enables users to interact with the system through kernel modules, while the AI layer handles I/O, device, and system components as if it were part of a robotic intelligence structure.
Tasks that normally require manual effort are handled quickly and efficiently by the AI system, allowing smoother and more autonomous interactions with your Linux environment.
Currently, the project supports core components such as I/O management, CPU control, and PCIe (PCIO) operations. The source code is open for contributors.
Support the project via PayPal:
👉 Donate on PayPal
Due to device support limitations, this project does not integrate with closed-source solutions such as nvidia-smi or similar proprietary APIs. Documentation for a proc-GPU support structure will be provided for contributors interested in extending GPU management.
Supported hardware currently includes Intel, AMD, and other integrated GPU and CPU components.
DE-AI is an AI management system designed specifically for desktop environments, providing limited AI support tailored to desktop use cases. It differs from Linux Kernel-AI, which offers full kernel-level AI control.
- Kernel-AI: Allows AI to take full control at the kernel level, enabling deep system integration and automation.
- DE-AI: Provides limited AI functionality within desktop environments, without full kernel integration.
Linux-AI provides two layers of AI control:
- ADMIN Layer: Grants full control to AI, carrying significant risk as permissions cannot be revoked. Rebooting is recommended after ADMIN-level operations.
- USER Layer: Allows limited, permission-based AI automation, safer for daily use.
Kernel-AI-cpud: Manages CPU resources in the background; requires ADMIN for full access.Kernel-AI-gpud: Controls GPU resource allocation and management; ADMIN recommended.Kernel-AI-gpm: Permission management tool for AI access levels.Kernel-AI-start: Starts AI services manually or via systemd; ADMIN recommended.
Linux-AI is primarily a kernel-level AI control system but can be installed alongside userspace components like DE-AI.
To install:
git clone https://github.com/Zamanhuseyinli/Linux-AI
cd Linux-AIBuild options:
- Using Python interpreter:
python Makemanifest.py- Using make:
make- Using CMake:
mkdir build
cd build
cmake ..
makeTwo build modes:
- Python-based interpreter version.
- Low-level kernel-integrated version via the
Makemanifesttool.
Specify user privilege level during build (e.g., Kernel-AI ADMIN).
⚠️ Warning:
Kernel-AI is under active development. It is strongly recommended to test it within a virtual machine first. The maintainers are not responsible for any misuse or illegal activity. Limited support for NVIDIA-SMI hardware is available. Report issues and request features on GitHub Issues.
The project is under development; some features may be incomplete. Testing in virtual machines or compiling as USER is advised until stable.
To install via AUR (currently only KDE Plasma DE-AI supported):
yay -S plasma-system-aiFor precompiled binaries:
yay -S plasma-system-ai-bin
