Startup ideas, the honest hub
Five proven frameworks for finding a startup idea, two topical guides for the highest-volume questions (SaaS and Reddit research), and a 30-day plan you can start today. Each section below links to a full deep dive. Or hand the whole thing to IdeaTwister and let 15 agents run all of them in parallel.
- Local-only
- No vendor lock-in
- One-time payment
- 7-day refund
The five frameworks
Every one of these has shipped real companies. The difference is which one fits you. Skim the summary, click into the deep dive that resonates, and come back to the others later.
Paul Graham
Live where the future already happened, then build what's missing. The most-cited essay on the topic for a reason - it works when you can name your frontier.
YC Requests for Startups
Y Combinator publishes a curated list of categories where partners want to fund founders. AI agents, vertical AI, robotics, on-device inference, and replacements for tools that broke in the LLM era.
Naval Ravikant
Specific knowledge times leverage. The thing you learned by accident, by obsession, or by working at one weird company for too long, multiplied by code, media, capital, or labor.
Camille Fournier
Productize the boring pain in the work you already do. Internal tools, ops workflows, platform pain. Boring here means has a buyer. CTOs already have line items in their budget for it.
Pieter Levels
Pick a tribe, build in public, ship in weeks. The tribe is the constant, the product is the variable. Anchor on a specific group of people you already belong to and ship the next obvious utility for them.
Topical guides
Two questions search engines see thousands of times a month. Each one gets a framework-agnostic deep dive.
How to find a SaaS idea
The specific patterns that produce SaaS ideas with real budget. Spreadsheet workflows, paid Zapier chains, and the phrase "we just emailed it back and forth."
Read the deep dive →Finding startup ideas on Reddit
Reddit and Indie Hackers are excellent signal sources if you treat them like data, not a brainstorm partner. The exact research method that turns complaint threads into a shortlist.
Read the deep dive →If you only have time for one
Three honest questions narrow it down quickly.
- 01Do you already inhabit a frontier? Yes → Paul Graham.
- 02Venture-scale or lifestyle-scale? Venture → YC RFS. Lifestyle → Pieter Levels.
- 03Specific knowledge accumulated? Lots → Naval. Working engineer with ops pain → Camille Fournier.
Still unsure? That is the case for the parallel approach. Hand your seed to IdeaTwister. Let 15 angles argue with each other for an hour. Read the ranked output. The choice between frameworks resolves itself when you can see all of them at once.
A concrete 30-day plan
Most founders do not fail at finding ideas. They fail at finishing the search. The fix is a fixed-date deadline. This plan fits into evenings and weekends without quitting your day job.
- D 1–3Write your seed sentence. Keep it under 30 words. Run it through the validation checklist to surface obvious gaps before spending compute.
- D 4–5Run IdeaTwister on the seed. About an hour, 50+ scored variations. Walk away during the run. Come back to a finished report.
- D 6–10Pick the top three by composite score. Validate the budget. Do people already pay anyone today to solve a worse version of this? If yes, keep going.
- D 11–20Talk to five potential buyers per remaining variation. Ten-minute calls, structured questions. How do you solve this today? What does it cost you? What would you pay for a better version? Listen, do not pitch.
- D 21–30Pick the variation with the cleanest yes. Not your favorite. Build the smallest useful artifact. Charge for it. The 30-day validation plan in the IdeaTwister report tells you exactly what to ship.
Tools and further reading
Idea validation checklist
The buyer-price-channel test, in one printable page.
Run IdeaTwister in 5 minutes
Install, run, read the report. No signup, no cloud.
The 5-dimension scoring rubric
How every variation is graded across Buyer Urgency, Market Proof, Solo Executable, Revenue Speed, and Defensibility.
Latest essays and case studies
Real seeds, real outputs, real lessons.
Frequently asked questions
How do most successful founders actually find their startup idea?
+
Almost none of them sit down and brainstorm cold. The pattern is consistent across YC, indie hacker, and venture studies. Founders find ideas by living inside a problem long enough to see a sharp, specific gap. Then they pressure test that gap against signals like what people pay for, what they hack together, and what they keep complaining about. Every framework on this hub points at the same thing. You don't invent ideas. You notice them.
Which framework should I start with?
+
If you already inhabit a frontier (AI agents, a niche subculture, a new platform), start with Paul Graham. If you want venture-scale, start with YC RFS. If you have an unusual career or hobby intersection, start with Naval. If you're a working engineer with operational pain, start with Camille Fournier. Earlier in your career, start with Pieter Levels so the tribe gives you what your biography hasn't yet.
How long should it take to find a startup idea worth pursuing?
+
If you work at it methodically, two to four weeks usually surfaces a small set of viable directions. The trap is spending six months researching because committing feels scary. The fix is forcing yourself to a written, scored shortlist by a fixed date. IdeaTwister compresses the synthesis step from weeks to about an hour by running 15 angles in parallel against your seed idea.
Should I pick a startup idea I am passionate about, or one with the biggest market?
+
Neither extreme works. Passion without a buyer turns into a hobby. A huge market without any personal investment turns into a slog you'll quit at month four. The honest answer sits in the middle. Pick a real, paying market where you have unfair access from a past job, a niche community, or a technical edge that lets you ship faster or sell warmer than a generalist.
How is IdeaTwister different from a typical startup idea generator?
+
Most generators output a list of unscored, undifferentiated suggestions written by a single LLM call. Clever, but commercially useless. IdeaTwister runs 15 specialized agents in parallel (Customer Segments, Pricing, Adjacent Niches, GTM Channels, Defensibility, and more), grounds each one in live web research, and scores every variation across five commercial dimensions. You walk away with a ranked, filterable shortlist instead of a brainstorm dump.
Five frameworks. One hour. Fifty scored ideas.
Hand IdeaTwister your roughest seed. Walk away. Come back to a ranked, filterable HTML report with $100K math and 30-day plans for the top five. $39 once. Runs locally. Yours forever.