Product Requirements Document
The one-page brief that tells you and your agent what you're building.
Copy →A free curriculum that takes you from idea to shipped software, no CS degree required. Internet Menace, again.
You don't have to read all of it. Read these five and you'll be functional by the weekend.
The shortest possible explanation of what this curriculum is and why it might be for you.
~15 min read02The five things that have to be true for AI-assisted building to actually work.
~20 min read03The minimum software setup that gets you from zero to coding-with-an-agent.
~45 min, half coffee break04A worked example: turn a vague idea into a buildable spec your agent can use.
~30 min write05Take the spec from step 04, hand it to the agent, push the result live.
~2 hrs buildWe’ve re-sequenced the curriculum six different ways, depending on who you are and what you want to ship. Pick the one that sounds most like you.
Choose your path →Or skip the path altogether. The wiki, the templates, and the glossary are all open. Search anything; start anywhere.
Grab one of these, fill in your specifics, hand it to your agent.
The one-page brief that tells you and your agent what you're building.
Copy →The project-context file every coding-agent project needs on day one.
Copy →Five pages, a contact form, a domain. Live by Sunday.
Copy →An agent that reads your inbox, drafts replies, and books meetings on your calendar.
Copy →Customers book themselves; the agent confirms, reschedules, and reminds.
Copy →Recent additions to the curriculum, the templates, and the glossary.
Nine templates became fifty: payments, auth, email, analytics, deploys, data pipelines, integrations, and a bench of new agents. Every one is a markdown file you copy, fill in, and hand to your agent. Section 9.2 still applies: write the fifty-first and send it back.
The original 2024 curriculum is back, reorganized as sections 0.7 through 0.17. If you finished v1, it's a reunion; if you didn't, it's the 180-day foundation this whole site quietly assumes.
Eight opinionated stacks, from a weekend Vercel-and-Supabase startup to causal agent observability with PyRapide. Hand your agent a coherent set of tools instead of letting it improvise one.