| pick your distro, get ZFS on root
kldload — your platform, your way, free
Source

AI

Run inference, fine-tune models, and build AI-powered infrastructure tools — all on ZFS-backed storage with NVIDIA GPU passthrough. These guides cover the full stack from local LLM hosting to AI-driven operations.

Getting Started

Getting Started

Set up your first AI workstation on kldload — NVIDIA drivers, CUDA, and local LLM inference with Ollama or vLLM.

Train on Your Infra

Fine-tune models on your own infrastructure data — logs, metrics, and configurations become training material.

Voice & Vision

Speech-to-text, text-to-speech, and computer vision pipelines running locally on GPU with ZFS dataset management.

AI for Infrastructure

AI for ZFS

AI-assisted ZFS management — anomaly detection on pool health, predictive scrub scheduling, and smart compression tuning.

AI for eBPF

Combine eBPF observability with AI analysis — automated root cause detection and intelligent alerting from kernel traces.

AI for WireGuard

AI-driven network management — traffic pattern analysis, tunnel health prediction, and automated peer configuration.

AI for Docker

AI-optimized container workflows — image layer analysis, resource prediction, and intelligent scheduling on ZFS.

AI for KVM

AI-managed virtualization — VM right-sizing, migration planning, and GPU resource allocation across KVM hosts.

AI for Kubernetes

AI-enhanced Kubernetes operations — cluster autoscaling, pod placement optimization, and predictive failure detection.