π The guide
Who it is for
- Managers and decision-makers wondering whether local AI meets their confidentiality and cost constraints. Each article opens with a jargon-free summary.
- Developers and administrators discovering or deploying local AI. The body of the articles stays technical, and each term links to its glossary entry.
What it covers
| Part | What you will find | Where to start |
|---|---|---|
| Foundations | Why memory decides everything | The journey of a prompt |
| Hardware | PCs, Macs, workstations and GPU servers | Hardware index |
| Software stack | Engines, models, clusters, RAG | Choose your model |
| Blueprints | Four scenarios, from developer workstation to datacenter | Cost comparison |
| Agents and assistants | AIs that use your data or act for you | Overview |
| Implementation | Install, test, monitor, secure | Getting started with Ollama |
The guide has more than 120 pages per language, many of them glossary entries. It reads like a digital garden, by following the links. The index offers a path for beginners and a table for choosing a scenario.
How it is written
The pages are written by AI agents, under the direction of a human author who sets the plan and the editorial line. Four rules frame this work:
- No figure without a source. Every numeric value relies on a reliable, dated source. A video or forum may provide a lead, never a reference.
- Every change is reviewed. It goes through a pull request that the author merges personally.
- Dates are visible. The bottom of each page shows its last change and, when it has been audited, its last verification.
- Perishable information is tracked. Claims that will age are put on a list, with a date for checking them again.
The update process is detailed in How it works.
Licence and reuse
The content is dedicated to the public domain (CC0 1.0): you may copy, adapt and reuse it, including commercially, without asking for permission. The code of the scripts is under the MIT licence.
Sourced corrections are welcome, as an issue or pull request.
GitHub repository: DamienBecherini/ia-on-prem-vault