Validate demand before you build
Pull real search-demand signal for your idea, its category, and a competitor — programmatically, no screenshots — and get a plain read on whether interest is growing or dying before you commit a week to building.
Library
Every workflow and explainer, filterable.
Filter by role, type, difficulty, or category. Every entry is tagged with the Founding Four role that runs it, plus Foundations for the shared context every role reads.
Role
Type
Difficulty
Category
No entries match those filters. Try clearing one.
Run the idea filter before you hire the team
Four gates and two human checks that turn 'should I build this' into a verdict (Keep, Bench, or Scrap) before you write a single role file.
Claude Code vs. Cursor for a non-engineer founder
Cursor is a full code editor with AI built in; Claude Code is conversation-and-diff-first with no editor to learn. The right pick depends on whether you want to live in an IDE or not.
Claude vs. ChatGPT: one practical question, not a benchmark war
Skip the benchmark comparisons — the useful test is running your own actual, recurring task on both and judging the output against what you know to be true.
Cowork vs. no-code builders
No-code builders are for structured, persistent apps with a defined data model. Cowork is for conversational, judgment-heavy workflows. Most founders need both, for different jobs.
Free vs. paid Claude: what's actually worth paying for
A framework for deciding whether to pay, since exact plans and prices change: usage volume, whether you need Cowork or Claude Code, and how much context you regularly need in one conversation.
Draft a case study from one happy-customer call
Turns one good customer call into a real case-study draft — built from what they actually said, not invented quotes.
Chat vs. Cowork vs. Claude Code: which one, when
Three surfaces, one underlying model — the choice is about how much you want to review each step, not which one is smarter.
The role you're about to hire for — could Claude still cover it
Before writing a JD, a plain framework for whether the role is actually a hiring problem or a workflow you haven't built yet.
The first three hires AI-native founders delay
Three roles that traditionally get hired early but that a founder running real workflows on Claude can often cover solo for longer than expected.
Draft the fundraising FAQ before investors ask
Drafts honest answers to the hard questions before you're put on the spot in a room, using your real numbers and your real weak spots.
A lightweight CRM from a spreadsheet
Turns a plain spreadsheet of prospects into something that behaves like a CRM — stage tracking, follow-up flags, and a weekly summary — with no new tool to buy.
Turn ten user interviews into a ranked feature list
Hand Claude all your interview transcripts at once and make it count and rank requests — with the quotes cited so you can trust it.
Drafting the seed deck narrative
Drafts the actual narrative arc of a seed deck from your real story — not slide-by-slide copy, the through-line that makes the slides make sense.
The real signal it's finally time to hire
Not 'I'm busy' — the actual signal is a specific, recurring, judgment-heavy task that a workflow genuinely can't cover, showing up consistently for weeks.
The 30/60/90 plan for the person you just hired
Drafts a specific, honest first-90-days plan for a new hire, instead of a generic template that doesn't reflect what your company actually needs from them.
A changelog that writes itself from git history
Turns a week of commit history into a real, user-readable changelog entry — without you writing it from memory after the fact.
The founder-does-sales playbook, built on Claude
A repeatable weekly rhythm for doing your own sales — research, outreach, and follow-up — without a sales background or a sales team.
A term sheet in plain English
A term sheet is a non-binding summary of a deal's key terms — the fastest way to understand one is to have Claude translate the actual document into plain language, clause by clause.
What is a context window, and why your memory files fit inside it
The context window is everything Claude can see at once for a given response — your prompt, your files, the conversation so far. Nothing outside it exists to the model.
What is an LLM, in terms that actually matter to a founder
An LLM predicts likely next words from patterns in its training — which explains both why it's useful and exactly where it'll quietly go wrong.
What is inference (and why it's the only part you actually touch)
Training happens once, ahead of time, and you don't touch it. Inference is every single response Claude gives you — the part you actually interact with.
What is RAG, and do you actually need it
RAG means fetching the relevant piece of a large pile of documents before answering, instead of stuffing everything into context at once.
What is synthesis, and why founders are bad at it
Synthesis is turning a pile of scattered user feedback into a ranked, countable signal — the step founders skip because it's tedious, not because it's hard.
Find the feature request you've been ignoring for a quarter
Surfaces the request that keeps quietly recurring across support tickets, calls, and interviews — the one that's easy to dismiss individually but adds up.
How to write a prompt that actually works
The gap between a bad prompt and a good one is almost never cleverness — it's context, specificity, and telling Claude what you actually want back.
A metrics dashboard with no analyst
Turns raw exports from wherever your data actually lives into a real, readable metrics view — no analyst, no BI tool license required.
Screen a stack of résumés without reading all of them
Scores a pile of résumés against a rubric you wrote once, so you spend your limited hiring time on the shortlist instead of the whole stack.
What is a data room, and how to assemble one fast
A data room is the folder of documents serious investors expect before they'll move — assembling it before it's asked for saves the exact week you don't have during a raise.
Permission modes: how cautious to be with Claude Code
Claude Code has explicit permission modes, from asking before every edit to acting more autonomously — the right setting depends on what it's touching, not on how much you trust it in general.
Use Claude Code to fix the bug you can't find
Turns a vague, hard-to-reproduce bug into a real fix by having Claude Code investigate the actual codebase instead of guessing from a description.
Why you hand Claude the spreadsheet, not a description of it
The single most common way founders get a vague investor update: they describe their metrics in a sentence instead of pasting the real numbers in.
Onboarding emails as a Claude-run sequence
Drafts a full onboarding email sequence in your own voice, grounded in what new users actually need to succeed with your product.
What a 'skill' is and when to build one
A skill is a reusable instruction set you write once instead of re-explaining every time — the difference between a one-off prompt and a real tool.
Ship a landing page with Claude Code without knowing how to code
Ships a real, working landing page by describing what you want and reviewing Claude Code's actual changes — no prior coding experience required.
What hooks are (and why you probably don't need them yet)
A hook triggers an automatic action around something Claude Code does — mostly an engineering tool, worth knowing about even if you rarely reach for it directly.
Subagents, when a founder actually needs one
A subagent is a specialist Claude Code can delegate a piece of work to — useful for heavier research or multi-part tasks, rarely needed for a founder's everyday work.
What is a connector (and the MCP behind it)
A connector lets Claude read and act on a real tool you already use — Gmail, your calendar, a CRM — instead of you copy-pasting between tabs.
What is ICP, and how to actually write yours
ICP means the specific person who buys fast and stays — writing yours down is what makes every other outreach workflow actually work.
Wire up Stripe with Claude Code
Adds real payment collection to your product with Claude Code — with the review discipline real money handling actually deserves.
Draft your product roadmap from a pile of Notion notes
Turns scattered planning docs and half-finished notes into a real, prioritized roadmap you can actually show someone.
Reading your first Claude Code diff
A diff shows exactly what changed, line by line — red for removed, green for added. Learning to read one is the actual skill that makes Claude Code worth using over chat.
Ship a waitlist in an afternoon
Ships a real, working waitlist page with email capture in one sitting — no page builder, no template fighting.
What is Cowork, and what founders actually use it for
Cowork is for anything that needs to run more than once, or take several steps on its own — the middle ground between chat and Claude Code.
Spec-first building: why you write the doc before the code
A short spec written before you build catches the vague parts of an idea while they're still cheap to fix — before you or Claude have written a line of code.
The terminal commands a founder actually needs
Six commands cover almost everything a founder needs to navigate a terminal well enough to use Claude Code's CLI — the rest you can skip.
What is Claude Code, in plain founder English
Claude Code reads a folder on your machine and remembers it — the difference between a chat tab and something that actually edits your files.
Your first 30 minutes with Claude Code
A literal, minute-by-minute walkthrough of opening Claude Code for the first time and getting it to do something real.
What is a memory file, and why every operator should keep one
A memory file tells Claude the things it would otherwise ask you every time — what you're building, who your users are, how you write.