Local AI Moat

Local AI Tutorials

From zero to your own AI workstation.

Step-by-step playbooks for running Ollama, DeepSeek-R1, Llama, and full agentic workflows on your own hardware. Real shell commands, hardware picks, and the gotchas no generic blog post will tell you.

14+

Tutorials

21+

Hours of content

Start here

The four tutorials our readers actually finish, and use every day.

Install Ollama on Ubuntu 22.04 with NVIDIA GPU

OllamaBeginner

30 min

Install Ollama on Ubuntu 22.04 with NVIDIA GPU

Stand up a fully functional local LLM stack on Ubuntu 22.04, NVIDIA drivers, CUDA, Ollama, and your first model, in under 30 minutes.

Read tutorial 8 sections

Install Ollama on Windows 11 with WSL2 and CUDA

OllamaBeginner

45 min

Install Ollama on Windows 11 with WSL2 and CUDA

Run local LLMs on Windows 11 the right way, WSL2, the WSL CUDA toolkit, and Ollama with full GPU acceleration. Native installer covered too.

Read tutorial 8 sections

Build Your First Dedicated AI Server From Scratch

AI ServerIntermediate

4 hr

Build Your First Dedicated AI Server From Scratch

Pick parts, assemble the box, install Ubuntu Server headless, and serve LLMs to your home network. A four-hour project that pays back forever.

Read tutorial 9 sections

DeepSeekBeginner

30 min

Run DeepSeek-R1 Locally with Ollama

DeepSeek's reasoning model rivals GPT-4 on math and code. Run the distilled 7B/14B/32B variants on consumer hardware in 30 minutes.

Read tutorial 7 sections

All tutorials

Filter by category, difficulty, or topic.

Install Ollama on Ubuntu 22.04 with NVIDIA GPU

Stand up a fully functional local LLM stack on Ubuntu 22.04, NVIDIA drivers, CUDA, Ollama, and your first model, in under 30 minutes.

Install Ollama on Windows 11 with WSL2 and CUDA

Run local LLMs on Windows 11 the right way, WSL2, the WSL CUDA toolkit, and Ollama with full GPU acceleration. Native installer covered too.

Install Ollama on macOS with Apple Silicon

Apple Silicon is shockingly good at LLM inference. Get Ollama running with unified memory acceleration on any M-series Mac in 15 minutes.

Build Your First Dedicated AI Server From Scratch

Pick parts, assemble the box, install Ubuntu Server headless, and serve LLMs to your home network. A four-hour project that pays back forever.

Run DeepSeek-R1 Locally with Ollama

DeepSeek's reasoning model rivals GPT-4 on math and code. Run the distilled 7B/14B/32B variants on consumer hardware in 30 minutes.

Run Llama 3 70B on a 24GB Consumer GPU With Q4 Quantization

Squeeze a 70-billion-parameter model onto a single RTX 4090 by quantizing weights to 4 bits. Quality stays high, VRAM usage drops to 24 GB.

Multi-GPU Inference With vLLM and llama.cpp

Scale a single model across multiple GPUs with tensor parallelism. Configure vLLM and llama.cpp, fix PCIe bottlenecks, double your VRAM.

GPU Passthrough on Proxmox for AI VMs

Run multiple isolated AI environments on one host using Proxmox VE with PCIe passthrough. Bare-metal performance, virtual machine convenience.

Build a $1,500 AI Homelab From Used Parts

The full parts list, source guide, and assembly walkthrough for a 48 GB VRAM AI homelab on a $1,500 budget. Used 3090s, AM4, no compromises that matter.

RTX Optimization for LLM Inference: Undervolting, Fans, Power Limits

Squeeze 10-15% more performance from your RTX card with smarter power and thermal settings. Less heat, less noise, lower power bill.

Quantization Explained: Q4 vs Q5 vs Q8 vs FP16

What quantization actually does to a model, why Q4_K_M is the default sweet spot, and when to deviate. A practical reference for picking quants.

Build Local AI Agents With Ollama, LangChain, and Custom Tools

Wire your local LLM to web search, file system, and Python tools. Build a fully offline agent that can do real work, no API keys.

Self-Host Open WebUI as Your ChatGPT Replacement

A polished, multi-user, self-hosted ChatGPT-clone running entirely on your hardware. Docker, 5 minutes, no API keys.

Use the Ollama API From Your Apps

Skip the CLI, call Ollama directly from Python, JavaScript, or curl. Covers the native API, the OpenAI-compatible endpoint, streaming, and structured output.

45 min•9 sections

Read tutorial

Browse by topic

Nine deep verticals, each with its own playbooks.

Pair these tutorials with our tools

Plan the build, validate the fit, then follow the playbook.

Stay Ahead of the AI Curve

Get weekly AI hardware news, benchmark updates, and deals in your inbox. Founding-subscriber list, be one of the first.

&check; No spam&check; Weekly digest&check; Unsubscribe anytime

Local AI Tutorials

Start here

Install Ollama on Ubuntu 22.04 with NVIDIA GPU

Install Ollama on Windows 11 with WSL2 and CUDA

Build Your First Dedicated AI Server From Scratch

Run DeepSeek-R1 Locally with Ollama

All tutorials

Install Ollama on Ubuntu 22.04 with NVIDIA GPU

Install Ollama on Windows 11 with WSL2 and CUDA

Install Ollama on macOS with Apple Silicon

Build Your First Dedicated AI Server From Scratch

Run DeepSeek-R1 Locally with Ollama

Run Llama 3 70B on a 24GB Consumer GPU With Q4 Quantization

Multi-GPU Inference With vLLM and llama.cpp

GPU Passthrough on Proxmox for AI VMs

Build a $1,500 AI Homelab From Used Parts

RTX Optimization for LLM Inference: Undervolting, Fans, Power Limits

Quantization Explained: Q4 vs Q5 vs Q8 vs FP16

Build Local AI Agents With Ollama, LangChain, and Custom Tools

Self-Host Open WebUI as Your ChatGPT Replacement

Use the Ollama API From Your Apps

Browse by topic

Pair these tutorials with our tools

Best GPU for Local LLMs

LLM VRAM Calculator

AI PC Builder

Benchmark Database

Stay Ahead of the AI Curve