Free • self-guided • build-first • commercially aware
A complete learning path for turning messy real-world processes into reliable automated systems using APIs, workflow platforms, Python, databases, AI where appropriate, browser/document automation, agents, and production engineering.
The target is not “become an n8n user.” The target is an automation engineer who can own the outcome.
understand the process → design → build → break it → recover → operate → document → prove value
- Open START-HERE.md.
- Read the Self-Study Operating System once before Phase 0.
- Begin with Automation Thinking even if you have used workflow tools before.
- Use the matching executable challenge from LAB_MAP.md as you progress.
- Keep your own workflows/code/progress outside the canonical curriculum repository; start from PROGRESS_TEMPLATE.md.
- Do not sell a capability you cannot pass at the corresponding commercial gate.
| Phase | Outcome | Typical effort |
|---|---|---|
| 0 — Automation Thinking | map work, bottlenecks, handoffs, risks | ~1 week |
| 1 — HTTP, APIs & Webhooks | integrate unfamiliar APIs reliably | ~2 weeks |
| 2 — Workflow Engineering | modular workflows, errors, debugging, monitoring | 2–3 weeks |
| 3 — Python Automation | custom clients, services, testing, n8n↔Python systems | 3–4 weeks |
| 4 — Data & Persistent Systems | SQL, state, idempotency, queues, concurrency | 2–3 weeks |
| 5 — JavaScript for Automation | focused JS/async/custom workflow logic | 1–2 weeks |
| 6 — AI-Powered Automation | structured AI, RAG, evals, human review | 3–4 weeks |
| 7 — Business Automation Systems | client-grade architectures from vague requirements | 3–4 weeks |
| 8 — Browser & Document Automation | reliable browser/file/document pipelines | 2–4 weeks |
| 9 — Agentic Automation | bounded agents, tools, state, approvals, recovery | 3–5 weeks |
| 10 — Production Reliability & Security | deploy, monitor, secure, back up, recover | 3–4 weeks |
| 11 — Commercial Engineering | discover, scope, price, deliver, hand over, support | continuous later |
| 12 — Automation Products | turn repeated solutions into reusable systems | open-ended |
Full detail: ROADMAP.md · curriculum map: CURRICULUM_MAP.md · executable labs: LAB_MAP.md
Deterministic logic when possible → AI when ambiguity requires it → human approval when consequence or uncertainty demands it.
The repository contains native lessons, curated resources, integration exercises, failure-injection/debugging work, realistic business projects, assessments, mastery gates, production/security material, commercial-readiness gates, and a controlled executable lab layer.
The labs include local API failure simulation, idempotency debugging, persistent-state design, AI evaluation, browser fixtures, bounded-agent policy checks, and an actual production fault injector. They are deliberately stable so the core course does not depend on random external websites.
This repository validates its phase structure, local links, and executable lab smoke tests on both Linux and Windows through GitHub Actions.
Run the same checks locally:
python scripts/validate_academy.py
python scripts/run_smoke_tests.pyCommercial activity starts before the technical curriculum ends, but only after capability gates. The academy teaches problem discovery, scoping, demos, pricing, delivery boundaries, licensing/ownership, security responsibility, handover, support, case studies, and productization. It does not promise income on a calendar.
Use the same curriculum while keeping separate implementation/progress repositories. Build key exercises independently first, then review one another's architecture, failure handling, documentation, and alternative solutions.
The academy uses n8n as a primary workflow platform because it is practical and current, but concepts come first. You should be able to transfer API, state, reliability, orchestration, and business-process skills to other platforms.
Educational content is CC BY 4.0; code examples are MIT-licensed. See LICENSE.md and CONTRIBUTING.md.
Start here: START-HERE.md