Course · step 2 of the From chat to agent course · 8 lessons
Your Jarvis: an assistant with the memory and personality you define
You learn the seven parts of an AI assistant, write who it is in four text files, ask a coding agent to build the core with tests, and connect it to Telegram. No programming skills needed: you decide, the agent builds, and the tests confirm it works.
- 8lessons
- 7parts of an agent
- 18hands-on exercises
- 3videos
The lessons
This is the deep dive for step 2. The overview of all four steps is in the From chat to agent course.
Lesson 1 · 15 minThe seven parts of a JarvisEvery agent, from the simplest to the most expensive, has the same seven parts: a door, a heart, an identity, a brain, a memory, hands, and a vault. Knowing these parts helps you choose better, spend less, and avoid getting locked into one product. INTELECTO is our example: small on purpose, so you can read through it.
Lesson 2 · 20 minDesign your Jarvis before you install itThe assistant becomes yours through text files, not code. Before installing anything, you write who it is, what it should and should never do, and what it needs to remember about you. This design answers the installation questions and becomes the four identity files.
Lesson 3 · 15 minThe gate: the Telegram bot, with one ownerTelegram is the simplest gate: you create the bot in two minutes and get a token. That token is the key to the gate. Whoever has the token controls the bot, and anyone can message a bot. That's why this lesson is half creation and half protection.
Lesson 4 · 20 minBuild without coding: you choose, the agent builds, the test confirmsYou don't write code. You choose the ingredients on a page, like at a grocery store, export a list, and ask Claude Code or Codex to build it by following the INTELECTO construction plan. Then you check it with tests and a conversation in the terminal, before connecting it to Telegram.
Lesson 5 · 15 minMemory: how it remembers you tomorrowMemory is the difference between an AI that forgets everything when it restarts and one that has known you for months. In INTELECTO, it’s a database on your computer, with keyword search and no API cost. Understanding how it searches changes how you tell it things.
Lesson 6 · 15 minThe brain: cloud or local, decided with your numbersThe AI model is the easiest part to swap. In the cloud, OpenRouter gives you access to more than a hundred models with one key. On your computer, Ollama runs at no cost per message and sends nothing outside your machine. The choice isn't a matter of taste: it's a table with speed, quality, and cost.
Lesson 7 · 15 minThe hands and the vault: tools with limitsWith tools, the assistant can read and write files, run commands, and save memories. An agent with hands needs limits. The INTELECTO vault has four: encrypted secrets, a list of forbidden commands, a time limit, and a log of everything that was run.
Lesson 8 · 15 minA real routine and Jarvis up and runningAn assistant you don’t use is no use. In this lesson, you turn a weekly task into a standing request, use it five days in a row, and note what to improve. Then you see the steps to keep it running: on your computer, with automatic restarts, and on a rented server, up 24/7.
Who this is for
- 1Anyone who has completed step 1 (terminal, folders, and .env) and wants their own assistant, not a chat website.
- 2Anyone who wants to understand what’s inside an agent before trusting it with their routine.
- 3Anyone who doesn’t code but is willing to ask a coding agent and check its work with tests.
What you’ll be able to do
- 1Explain the seven parts of an agent and where each one lives in INTELECTO.
- 2Write the SOUL, AGENTS, USER, and MEMORY files for your assistant.
- 3Build the core with Claude Code or Codex and verify it with tests.
- 4Connect it to Telegram with only one owner, choose its brain using your own numbers, and follow a routine for a week.