Build an AI OS without coding
The three parts (model, memory, hands), the four Cs, the onboarding dossier, recipes, routines and the morning dashboard.
The problem
Your assistant forgets you every morning.
The model thinks well. What's missing is the house around it: memory, identity, tools and limits. Each course in the training covers these parts.
- Every conversation starts from scratch and you repeat who you are, what you do and how you like things.
- The assistant lives on a website: it doesn't talk to you on Telegram or touch your files.
- You're stuck with one product because you don't know what parts it's made of.
- Giving tools to the agent seems risky, and you don't know where to set the limit.
- Your data goes outside because you don't know how to run a model on your computer.
After the training
By the end of the 6 courses, you will be able to:
Build your Jarvis with INTELECTO
Identity, channels, memory, zero-trust security against prompt injection, skills and deployment with monitoring.
Make the assistant improve with use
The eight elements, the five garden files, the failure log and autonomy that rises with evidence.
Put Jarvis on Telegram
openpcbot v3 installed and operated: the path of a message, brain and connectors, agents, skills and cost.
Run everything on your own computer
Ollama, the right model for your hardware, the 100% local Hermes Agent, and background agents running 24/7.
Design an AI operating system
The 6 layers, identity, rules and hooks, skills and agents, with sales, support, and content OSes.
6 courses in order · 84 lessons and modules
Six courses, from the no-code assistant to the system in production.
Courses 1 and 2 give you the foundation and your first Jarvis. Course 3 teaches you to make it improve. Courses 4, 5, and 6 are paths to go deeper: Telegram, local model, and system architecture.
- 1
Beginner
Build your AI OS
The assistant that doesn't forget: the three pieces, the four Cs, the dossier, connectors, recipes, routines, a wiki with a daily dashboard, and your 30-day plan.
9 lessonsOpen course 1 - 2
Your Jarvis
INTELECTO Course
Six tracks from what INTELECTO is to deployment: identity, channels, memory, zero-trust security, providers, skills, webhooks, and planning your assistant.
18 modulesOpen course 2 - 3
Cultivationv7.2
Cultivated AI
The eight elements and the five files that cultivate, the failure log, and evidence-based autonomy, applied to your Jarvis.
8 lessonsOpen course 3 - 4
Telegram
openpcbot v3 — your local Jarvis
The path of a message, starting up and operating v3, commands, brain and connectors, conversation control, agents, skills, and cost.
8 modulesOpen course 4 - 5
100% local
Agente Hermes + Ollama
Ollama installed, the right model for your hardware, and local Hermes across seven projects: private chat, memory, GitHub, 24/7 agents, and your phone.
19 modulesOpen course 5 - 6
Advanced
AI Operating System Engineering
The 6 layers of an AI OS, the step-by-step with /os-coach, and real domains: sales, support, content, consulting, and the OS that works for hours on its own.
22 modulesOpen course 6
Who it's for
Made for you if
- You want an assistant that is yours alone, that remembers you and talks to you on Telegram or WhatsApp.
- You would rather understand the pieces than be stuck with a ready-made product.
- You want to run AI on your own computer, without sending your data out.
- You already use Claude Code or Codex and want to turn it into your work system.
Not for you yet if
- You haven't left the chat yet: start with the From chat to agent area.
- You want an assistant that works without you having to define what you expect from it.
Included in the training
The courses come with ready-to-install Jarvis assistants.
The program builds on open INEMA projects on GitHub, with a guide in three languages. They come along, together with the videos on the topic.
Projects and kits
Videos
Why this training
What changes compared to piecing together loose tutorials.
The same seven pieces in every course
Channel, brain, identity, memory, tools, loop, and security. You compare systems and choose without depending on a single product.
From no-code to advanced
Course 1 doesn't require programming. The following ones take you to the terminal, the local model, and the system architecture.
Projects you install today
INTELECTO, Jarvis v7, Jarvis Modelo, openpcbot v3, Astra 2Cérebro, and INEMA Agent Runtime, all open.
Three languages, same content
All six courses are available in Portuguese, English, and Spanish.
Questions
Before you start.
Do I need to know how to code?
For course 1, no. In INTELECTO you choose the pieces and Claude Code or Codex writes the code with tests.
Do I need to follow the order?
Start with 1 and 2. After 3, choose: Telegram in 4, local model in 5, or system architecture in 6.
Which tools do I use?
Claude Code or Codex in the terminal, Telegram or WhatsApp as the channel, models through OpenRouter or local ones through Ollama.
How long does it take?
That's 84 lessons and modules: 9 in course 1, 18 in INTELECTO, 8 in Cultivated AI, 8 in openpcbot v3, 19 in local Hermes, and 22 in AI OS engineering.
Does my data leave my computer?
It depends on the brain you choose. Course 5 shows how to run everything locally with Ollama, sending nothing out.
A week from now, you could have an assistant all your own.
Start with course 1. In 9 lessons you build your AI OS without programming and leave with the 30-day plan.