$39
Specification for AI
Write specs that an AI can't misread. PRD-as-tests, acceptance criteria, deterministic checks.
Try Lite freeBack to all simulators
Simulator
Agent for Human
$19Lite playable · full track on waitlist
7-day course: hire your AI. From SOUL.md to a real working colleague — with daily exercises auto-graded by FolderAI.
Free · 12 runs/hour · streams in 5-10 sec
Want the full track ($19)? Get notified when it ships.
The problem this solves
You tried 'AI employees' off Twitter and came back disappointed. Every Monday you re-explain the project because the agent forgot last week. You'd let it send emails — if you trusted it not to send the wrong one. So it stays a glorified autocomplete you babysit, and the colleague you were promised never shows up. The problem isn't the model; it's that nobody ever gave the agent an identity, a memory, or a way to earn your trust.
After completing this
In seven days you stand up a working AI colleague — not a chat history, a portable spec. It shows up Monday knowing who it is, who you are, and what's off-limits. It remembers last week. It earns autonomy step by step up a trust ladder you control, from read-only to acting on its own. You walk out able to hand that spec to someone else, or sell it, instead of restarting a chat.
How it works
It's a simulator, not an article
You don't read about AI colleagues — you build one, with daily exercises auto-graded by FolderAI instead of lectures.
You answer first, theory after
Each day you make the call — write the SOUL, set the trust rung — before the why. The reasoning lands because you already acted.
From the demo, a direct bridge to the paid track
The free Lite playground flows straight into Day 1 of the paid track, where SOUL.md becomes a colleague you ship by Day 7.
Who this is for
Solo founders who keep restarting a chat to re-feed context
Stand up one agent that survives Monday morning without a briefing — SOUL.md and three memory layers carry context between sessions.
Knowledge workers ready to delegate but scared to lose control
Calibrate a trust ladder you defend to yourself: read-only first, autonomous only on actions you've already watched ten times.
Consultants packaging themselves into an assistant for clients
Ship a FolderAI-validated, portable spec you can hand off or sell — the artifact, not the conversation.
Not for you if
- People who want a no-code 'AI employee' button — you'll write SOUL.md and wire memory yourself, by design.
- Anyone unwilling to put the agent on a leash first; the trust ladder is the whole point.
- Teams who already run identity + memory + approval-queue patterns in production.
When you'll reach for this
01
You want one agent that survives Monday morning without a refresher
Day 1–2 walks you through SOUL.md and IDENTITY.md so the agent shows up knowing who it is, who you are, and what's off-limits.
02
Your agent forgets last week and you keep re-explaining the project
Day 3–4 wires three memory layers + a PARA graph so context persists between sessions instead of leaking out of the prompt window.
03
You'd let the agent send emails — if you trusted it not to send the wrong one
Day 5–6 builds the approval queue and trust ladder. Read-only on Monday, autonomous on actions you've already watched ten times by Friday.
04
You want a spec you can hand off or sell, not a chat history
Day 7's final spec is FolderAI-validated and portable — the artifact, not the conversation.
Powers you walk away with
- Stand up a working AI colleague in a week, not a quarter
- Write a SOUL.md that survives a model swap without behavior drift
- Calibrate the trust ladder so the agent earns autonomy instead of demanding it
- Ship a spec other people can run — not a prompt only you understand
- Stop restarting chats every Monday morning
What's inside
- 01Day 1–2: SOUL.md & IDENTITY.md — voice, role, scope
- 02Day 3–4: 3-layer memory & PARA knowledge graph
- 03Day 5: Tools & approval queue
- 04Day 6: Trust ladder — from read-only to autonomous
- 05Day 7: Final spec, validated and shipped
Same SOUL.md/IDENTITY.md pattern every internal FolderAI agent ships with.
What stays in your portfolio
- A SOUL.md + IDENTITY.md that survives a model swap without behavior drift.
- A three-layer memory + PARA graph that keeps context alive between sessions.
- A calibrated trust ladder — read-only to autonomous — you can defend.
- A FolderAI-validated, portable agent spec others can run, not a chat log.
Pairs well with
$49
I'm Agent Creator
Build agent folders end-to-end: pipeline, sub-agents, memory layers, conflict detection.
Try Lite free
FAQ
Is the Lite demo free?
Yes. The Lite playground is free and needs no card. The full 7-day track is $19 (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 decide to take the full 7-day track after the demo.
How long does it take?
The Lite demo is one sitting. The full track is seven days of short daily exercises, auto-graded by FolderAI — most people do one a day.
What's next after the demo?
The Lite demo bridges straight into Day 1 of the paid track — SOUL.md and IDENTITY.md — then memory, tools, the trust ladder, and a shipped spec by Day 7.
Do I need to code?
No. You write markdown identity and memory files, not code. If you can write a clear job description, you can do this.
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.