Loading...

The Agent Foundry

All-access pass
to agentic playbooks.

Get instant access to playbooks and repositories for shipping production-grade systems: 6 implementation-focused books, deployable reference repositories, and managed-service-first architectures for application backends, authentication and tenant isolation, durable background jobs, event-driven lifecycle workflows, subscription and usage-based billing, rate limiting, retries, and production observability.

[ ++ ]
WHAT READERS SAY

Feedback from readers of the playbooks and write-ups these books and repositories are built on.

Really enjoyed this, especially the emphasis on system design over model choice. I shared it in a LinkedIn post as it aligns closely with how I've been thinking about inference orchestration. Appreciate you putting this out.
Peter
This is crazy! With your exact llama.cpp config, performance on my RTX 4090 is insane for short prompts: input 5,386 tokens/s (no caching), output 165 tokens/s. Perceived output quality is excellent too, on par with gemini-3-flash for my use case. Thank you very much for sharing!
Skillconomy
Thank you! Massive effort and greatly appreciated. This is a great write-up and incredibly helpful!
Jeff
Congratulations! I finally found a resource that explains things simply, clearly, and directly, and, most importantly, focused on everyday use.
Pierre
I've read a hundred resources about this stuff and most of them feel like sales pitches. This was actually really helpful though!
Arjun
For real though, these should be way more popular!
Adam
This hit all the right notes
Kamrun
You put great value in this resource, thank you!!
ノウラ
Impressive breakdown of how efficient architectures are redefining what's possible on consumer hardware. This really highlights a meaningful shift for local AI development. The practical insights on optimization and real-world tradeoffs make this especially valuable for anyone building serious agentic workflows.
Olivia
Good to read expert opinion, keep it up!
Thomas
[ 01 ]
THE LIBRARY — GROWING

6 books shipped, and growing.

Open full catalog
Agentic SaaS Playbook in 2026 cover

Shipped

Agentic SaaS Playbook in 2026

Production blueprint for architecting, shipping, and operating agentic SaaS

Agentic Software Factories cover

Shipped

Agentic Software Factories

Turn coding agents into a governed production line: intent as versioned specs, an assembly line, quality control, coherence, provenance, and accountability — in current Claude Code and Codex syntax

AgentOps: The Playbook For Businesses cover

Shipped

AgentOps: The Playbook For Businesses

Operating model and architecture playbook for enterprise-grade agent systems

GPU Architecture and Programming cover

Shipped

GPU Architecture and Programming

Two volumes, one machine: the GPU platform from SIMT cores to GPU-cloud economics — with the math — then one CUDA kernel read line by line at four altitudes

OpenClaw Agents: Automate and Scale Operations cover

Shipped

OpenClaw Agents: Automate and Scale Operations

Implementation guide for production OpenClaw agent systems

Ultimate Guide to Local LLMs in 2026 cover

Shipped

Ultimate Guide to Local LLMs in 2026

Comprehensive technical reference for deploying, optimizing, and productionizing local LLMs

[ 02 ]
WHAT IT WOULD COST SEPARATELY
Agentic SaaS Playbook + SaaS boilerplate: all 7 planes (Runtime, Control, Memory, Data, Security, Observability, UX)€79
OpenClaw Agents Runbook: VPS hardening, deployment playbooks, skill integrations€49
AgentOps Book: enterprise operating model, governance, and production reliability playbooks€69
Ultimate Guide to Local LLMs in 2026: quantization, runtimes, hardware planning, deployment war stories€89
GPU Architecture and Programming: SIMT cores to GPU-cloud economics, plus one CUDA kernel read at four altitudes — with the math€99
Agentic Software Factories: turn coding agents into a governed production line — specs, quality control, provenance, accountability€79
Production SaaS templates & starter repos (auth, billing, tenant isolation, retries)€149

Standalone value

€613+

The Complete Library

€49.99once

Members get all of it — plus everything that ships next

Membership

The Agent Foundry

Freefor 7 days, then €14.99/mo

  • Everything in The Complete Library
  • Every future book on release
  • New playbooks, repos, and assets as models and stacks change
  • Direct access to the author

€0 today · cancel anytime

6 books shipped35+ premium modules€613+ standalone valueUpdated as stacks change
What you getLibraryMembership
Every book in today's library (6)
All code repositories & premium assets
Lifetime updates to the included books
New books as they ship
New playbooks, repos & assets as stacks change
Direct access to the author
Price€49.99 once7-day free trial, then €14.99/mo
[ 03 ]
WHY THIS EXISTS

The industry is full of impressive demos, workflow walkthroughs, and repositories. But a working prototype is not a production system. The real engineering begins at the point where most examples stop: authentication, authorization, tenant isolation, billing, durable execution, retries, observability, data ownership, and the failure modes that only appear under real usage.

The Agent Foundry is built for founders, engineers, and operators closing that gap. It provides implementation-level guidance, production-oriented reference architectures, and code shaped by actual product constraints—not idealized examples.

Every playbook is derived from building and operating agentic AI products in practice. As models, infrastructure, pricing, and platform capabilities evolve, the material evolves with them. The result is a continuously maintained engineering library for the systems teams usually have to discover the hard way: identity and access management, multi-tenant data boundaries, subscription and usage-based billing, background jobs, fault-tolerant workflows, retries, and end-to-end observability.

These are the architectures, implementation patterns, and repositories I would have wanted before shipping my first agentic product—designed to be inspected, adapted, deployed, and operated under real production conditions.

[ 04 ]
COMMON QUESTIONS
"One-time Library or Membership — which should I pick?"

If you want the books and code as they exist today, buy the Complete Library once (€49.99), it's yours forever, including future updates to those books. If you want the library as a living system, every new book on release, new repos and playbooks as the stack changes, and direct access. The membership is built for that. You can own the Library now and join the membership later, and they will stack.

"Do Library buyers get future books?"

No, and that's deliberate. The Library covers the books published at the time you buy, plus all future updates to those books. Brand-new books ship to members first. Join the membership anytime to get every new release.

"I can piece this together from free content."

These are integrated systems and execution paths, not disconnected blog posts, the premium planes (Data, Runtime, Control, Security, Observability) are 10x deeper than the free chapters, and the repos reflect real product constraints.

"I don't need another course."

Good, because this isn't one. Playbooks are engineering references: implementation patterns and production repos, updated as the ecosystem moves.

"What if it's not worth it?"

Everything here is reversible. The membership starts with a 7-day free trial, you can cancel before it ends and you're charged nothing. And there's a 7-day money-back guarantee on any purchase: email me and I'll refund you, no forms.

"I'm too early-stage for this."

That's the best time. You can install durable product foundations such as auth, billing, tenant isolation, observability, before they become expensive to retrofit.