No cloud · No subscription · No account

Run AI on hardware you own.

Four printed-quality PDF manuals that take you from your first local LLM to fine-tuned, air-gapped, fully private AI — written for people who don't code, on Windows, macOS and Linux.

Browse the manuals How delivery works

4 manuals50+ chapters total

40+ modelsLlama · Flux · Wan2.1 & more

3 platformsWindows / macOS / Linux

Pay onceFree lifetime updates

Why run AI locally

The cloud reads everything you type.

Every prompt you send to a hosted chatbot leaves your machine. Locally, the model is a file on your disk — nothing goes anywhere.

Privacy Your data never leaves Contracts, medical notes, client work, personal journals — processed entirely on your own CPU or GPU. Works offline, on a plane, or air-gapped.

Cost £0 / month, forever No API credits, no per-token billing, no £20/month plan. Generate millions of tokens and thousands of images for the price of the electricity.

Control Every parameter is yours Swap models, tune temperature and context, build custom personas, fine-tune on your own data. No rate limits, no content filters you didn't choose.

The manuals

Four manuals. One payment each.

Each is a standalone, bookmarked PDF with every command written out in full — no video to scrub through, nothing left ambiguous. Swipe through the series.

Manual 01 / 04PDF · Instant

Starter

First Boot

Your first local LLM talking in under an hour.

Chapters	13 · incl. 25 errors solved
Covers	Ollama · Llama 3 · Mistral · Phi-4 · Open WebUI
Hardware	Any — CPU-only setups included

Ollama setup on Windows, macOS & Linux
ChatGPT-style interface, fully offline
Hardware tiers & GPU compatibility tables
Prompt engineering for local models
Image generation
Video & fine-tuning

£15one-time · +5% delivery

Manual 02 / 04PDF · Instant

Builder

Image Gen

Unlimited AI images on your own GPU.

Chapters	15 · incl. 20 problems fixed
Covers	A1111 · ComfyUI · SDXL · Flux.1 · ControlNet
Hardware	NVIDIA 6 GB VRAM min. recommended

AUTOMATIC1111 & ComfyUI, step by step
Flux.1 — including 8 GB VRAM setups
ControlNet pose & composition control
LoRA styles, Civitai model management
Upscaling, inpainting & img2img
Video & fine-tuning

£20one-time · +5% delivery

Manual 03 / 04Most bought

Creator

Full Stack

Text, image and video — the complete local creative suite.

Chapters	14 across text, image & video
Covers	Wan2.1 · CogVideoX · LM Studio · llama.cpp · DeepSeek-R1
Hardware	Video from 8 GB VRAM; 24 GB for 720p

Wan2.1 text-to-video, full install & prompting
CogVideoX image-to-video & AnimateDiff
LM Studio, llama.cpp & quantisation picks
Qwen2.5-Coder in VS Code, locally
LiteLLM multi-model routing
Fine-tuning

£30one-time · +5% delivery

Manual 04 / 04PDF · Instant

Ultimate

Total Control

Fine-tune, quantise, RAG, voice and agents — full model ownership.

Chapters	15 · deepest of the four
Covers	Unsloth · QLoRA · GGUF · ChromaDB · Whisper · AutoGen
Extra	Private Discord community access

Fine-tuning with Unsloth & Axolotl
Dataset preparation & QLoRA hyperparameters
RAG pipelines: chat with your own documents
Whisper + Kokoro fully offline voice assistant
AutoGen & CrewAI local agent systems
Security hardening & air-gapping

£40one-time · +5% delivery

Tier select · drag Unit 01/04

Payment & delivery

Verified on our server. Delivered in seconds.

No account to create. Your download unlocks only after PayPal confirms the payment to our server — and a backup link goes to your email.

Step 1 — Pay

PayPal or any card

Checkout opens right on this page. Pay with your PayPal balance or any debit/credit card — a PayPal account isn't required.

Step 2 — Verify

Server confirms it

Our server checks the transaction directly with PayPal — amount, currency and status — before anything unlocks. No client-side tricks.

Step 3 — Download

PDF + email backup

A secure download link appears instantly and is also emailed to your PayPal address. The link stays valid for 90 days — after that, email us and we'll reissue it.

Questions

Before you buy.

No. The Starter manual includes CPU-only setups that run on a 16 GB RAM laptop with no graphics card. Image and video generation do want an NVIDIA GPU — every manual opens with an honest hardware table so you know exactly what your machine can do before you spend a minute installing anything.

Yes — that's who these are written for. Every command is given in full, in order, with the expected output shown. Each manual ends with a troubleshooting chapter covering the exact errors people actually hit, and every purchase includes email support with a 24-hour response time.

Each manual is standalone and assumes the knowledge of the ones before it. First Boot gets your text AI running; Image Gen adds Stable Diffusion and Flux; Full Stack adds video and advanced text tooling; Total Control covers fine-tuning, RAG, voice and agents. Buy the one that matches where you want to end up.

Local AI tools move fast, so the manuals get revised as tools and models change. When a manual is updated, you can re-download the latest version with your original link — or email us with your PayPal receipt and we'll reissue it. No extra charge, ever.

Because every manual is a digital file delivered instantly, purchases are non-refundable. That said, you're not on your own — if anything doesn't work on your setup, email hhcommunications@proton.me with your OS, GPU model and the exact error, and we'll work through it with you until you're running. Support is included with every manual, and we reply within 24 hours.

Check the email PayPal has on file — a backup link is sent there automatically at purchase. If it's expired or missing, forward your PayPal receipt to hhcommunications@proton.me and we'll send a fresh link within 24 hours.

Fair use

Manuals are licensed for personal use. Don't redistribute or resell the PDFs — we're a small operation and every copy sold funds the next revision.

Support

Stuck on any step? Email hhcommunications@proton.me with your OS, GPU model and the exact error message. Replies within 24 hours, included with every manual.