$19
Agent for Human
7-day course: hire your AI. From SOUL.md to a real working colleague — with daily exercises auto-graded by FolderAI.
Try Lite freeBack to all simulators
Simulator
I'm Agent Creator
$49Lite playable · full track on waitlist
Build agent folders end-to-end: pipeline, sub-agents, memory layers, conflict detection.
Free · 12 runs/hour · streams in 5-10 sec
Want the full track ($49)? Get notified when it ships.
The problem this solves
Your one-prompt agent grew into a 4,000-token monster nobody can edit. You want sub-agents, but every example online deadlocks at scale. The agent loses the plot after step 3 of a long task because everything's crammed into one system prompt. You've hit the wall where prompt-stuffing stops working and you don't have an architecture to replace it — so you keep patching the monolith and hoping.
After completing this
You decompose any task into an 8-stage pipeline without overengineering — each stage testable, each handoff explicit, each prompt short enough to read. You decide spawn-vs-merge for sub-agents by the work, not the trend, and wire three memory layers without inventing a database. You catch conflicts before they ship a Frankenstein output, and publish a folder others can fork and credit you for.
How it works
It's a simulator, not an article
You don't read about agent architecture — you design a real pipeline, spawn sub-agents, and watch the conflicts surface before they ship.
You answer first, theory after
You make the spawn-or-merge call before the explanation. The pattern sticks because you committed to a decision first.
From the demo, a direct bridge to the paid track
The free Lite playground flows into the full track, whose capstone puts your agent in the public library with forkable lineage.
Who this is for
Builders who shipped a single-prompt agent and hit the wall
Split the monolith into 8 testable stages, so the agent stops losing the plot after step 3 of a long task.
Indie devs assembling their first multi-agent product
Learn spawn-vs-merge with the failure modes shown, not hidden — so your sub-agents don't deadlock the moment they scale.
Tech leads choosing between LangGraph, CrewAI, and rolling their own
Get the architecture decisions framed by the work, not the framework hype — and a 3-layer memory pattern you own.
Not for you if
- People who've never shipped even a single-prompt agent — start with 'Agent for Human' first.
- Anyone hunting for the one framework that solves everything; this teaches patterns, not allegiance.
- Teams already running tested multi-stage pipelines with conflict detection in production.
When you'll reach for this
01
Your one-prompt agent is now a 4,000-token monster nobody can edit
Pipeline design splits the monolith into 8 stages — each stage testable, each handoff explicit, each prompt small enough to read.
02
You want sub-agents but every example online deadlocks at scale
Sub-agent patterns cover when to spawn, when to merge, and when the parent should just do it itself. With the failure modes shown, not hidden.
03
Your agent loses the plot after step 3 of a long task
Three memory layers (working, episodic, semantic) keep state where it belongs instead of stuffing everything into the system prompt.
04
You want your folder in a public library with credit and lineage
Final project ships to the FolderAI public library, with version history and a forkable lineage — your name on it, your patterns reusable.
Powers you walk away with
- Decompose any task into an 8-stage pipeline without overengineering
- Decide spawn-vs-merge for sub-agents based on the work, not the trend
- Wire 3-layer memory into a folder without inventing a database
- Detect conflicts between sub-agents before they ship a Frankenstein output
- Publish a folder others can fork and credit you for
What's inside
- 01Design 8-stage pipelines (QUIZ → EXECUTION)
- 02Sub-agent patterns: when to spawn, when to merge
- 033-layer memory in every folder you ship
- 04Final project: your agent in the public library, with lineage
Every public folder on FolderAI is built on the same 8-stage pipeline this course teaches.
What stays in your portfolio
- The ability to decompose any task into an 8-stage pipeline without overengineering.
- A spawn-vs-merge decision rule for sub-agents based on the work, not the trend.
- A 3-layer memory pattern you can wire into a folder without a database.
- A published, forkable folder in the public library with your lineage on it.
Pairs well with
$99Ch1 free
Automation AI Engineer
Ship like a team of one. Ralph-loop, TDD prompts, parallel worktrees, self-healing pipelines.
Start Ch1 free
FAQ
Is the Lite demo free?
Yes. The Lite playground is free and needs no card. The full track is $49, or reserve a seat now to lock launch pricing.
Do I need a card to try it?
No card for Lite. You only pay if you take the full track after the demo.
How long does it take?
The Lite demo is one sitting. The full track is self-paced and ends with a real capstone: your agent in the public library, with lineage.
What's next after the demo?
The Lite demo bridges into the full 'I'm Agent Creator' track — 8-stage pipelines, sub-agent patterns, 3-layer memory, and a published folder with your name on it.
Do I need a specific framework?
No. The patterns are framework-agnostic — you can apply them in LangGraph, CrewAI, or a folder you roll yourself.
Start with the free part
No card, no commitment — try it first, decide later.
Try Lite free — no signupMade by mihalkevich.com — runs on the FolderAIThe engine that powers the courses and simulators engine.