Understand AI and use it well
- Understand models, prompts and workflows
- Complete browser-based exercises
- Know where AI can help
GEMDUCK · AI PROJECT LAB
This is not another theory course. Download the course folder and open it in Codex or Claude Cowork. The AI shows you the finished result first, then guides one step at a time until your project runs and passes acceptance.
WHY THIS COMES AFTER THE FREE COURSE
PREVIEW BEFORE YOU BUY
AI WORKFLOW SPECIALIST
Confirm that you opened the course root folder and that the required tools are available.
Open the reference so you know exactly what you are building.
The AI explains starter and helps change your name and positioning—without dumping ten commands on you.
The AI interviews you about experience, skills and contact details, then writes approved content into the page.
Open the page locally. If something fails, diagnose it from the files and actual output.
Check mobile layout, links and content, then record the evidence in PROJECT-REPORT.
NOT A FOLDER OF PROMPTS
WHAT YOU WILL ACTUALLY BUILD
WEEKLY ACQUISITION REPORT
16 COMPLETE PROJECTS
Build a mobile-friendly personal brand website with safe course samples, pass local checks, then decide whether to publish.
Create a lead-generation page for a real service with pricing, FAQs and an enquiry form.
Turn a repeated calculation into a browser tool with inputs, rules, results and a copy button.
Build a browser game with start, scoring, increasing difficulty and replay.
Read test emails, classify and prioritise them, then create reply drafts that wait for human approval.
Turn messy meeting notes into decisions, owners, deadlines and questions that still need confirmation.
Extract a batch of text invoices into an auditable CSV and flag missing or duplicate records.
Answer questions about a set of local documents with filenames and line-number citations.
Answer from approved FAQs with citations and escalate low-confidence questions to a person.
Turn one original long-form article into email, LinkedIn and short-video drafts with human review.
Score and rank form leads with visible rules, then recommend a next step.
Build a dated competitor report from saved primary material with links, evidence levels and explicit unknowns.
Turn a CSV into auditable KPIs, trend charts, an anomaly list and a standalone HTML report.
Fetch, deduplicate and rank RSS items, then create a daily brief with original links.
Turn a precise trading hypothesis into a historical backtest and compare it with buy and hold.
Start with one narrow problem and build a SaaS prototype with demo login, core value, usage limits and a pricing page.
HOW IT WORKS
After purchase, get the complete ZIP from your private download centre.
Codex Desktop or Claude Cowork works best. Other folder-aware AI coding tools may also work.
The AI reads START-HERE, the tutor protocol and your progress, then advances one checkpoint.
Every project must run, pass tests and record evidence in PROJECT-REPORT.
PROJECT LAB IS FOR YOU IF
IT IS NOT THE RIGHT FIT IF
Best: Codex Desktop, Claude Cowork.
Works: Claude Code and folder-aware AI coding tools.
Limited: web chats that accept only single-file uploads.
The course pack itself calls no paid model API. Your cost depends on the subscription or usage allowance of the AI tool you choose. Every reference starts with local sample data.
Email stays in draft; quant projects use historical backtests only; research keeps sources; support, finance and high-impact decisions require human review.
QUESTIONS BEFORE YOU BUY
No. It is 16 local project folders. Each includes a finished reference, starter, AI tutor instructions, checkpoints, tests and an acceptance report. Your AI tool reads those files and guides the hands-on work.
You can, but beginners usually lack a precise goal, runnable reference, correct order and acceptance standard. Project Lab fixes those inside the course files so the AI cannot start by generating a pile of work you cannot judge.
Yes. The first projects begin with finished HTML files that open in a browser, and the AI advances one step at a time. You still need Codex Desktop or Claude Cowork; the private download page guides installation, extraction and your first message.
You enter a private download centre: download the ZIP, choose the Codex or Claude Cowork setup guide, open the course root folder and copy “Start the course”. You will not be left guessing after download.
The folder itself calls no paid API. Model cost depends on the AI tool and plan you choose. Every project starts with local sample data and clearly labels any external service.
The digital course has a seven-day refund period under the terms. It does not promise income, employment or investment results.
V1.0 · ONE-TIME PURCHASE