The framework from "How to Get and Evaluate Startup Ideas", the YC Startup School talk by Jared Friedman, embedded above. His recipes became the form's questions; his criteria became the score on each idea. We are not affiliated with Y Combinator, we just think it is the best public material on the subject.
Free Startup Idea GeneratorBuilt on Y Combinator's method for finding ideas
Most idea generators hand you a list anyone could have. This one starts from your skills and your experience, then scores each idea against the questions YC asks. Based on Jared Friedman's Startup School talk, below.
First generation freeConfidentialBased on the YC method
Tell us what you bring
The more concrete you are, the more specific your ideas will be. Skip the buzzwords and write like you would to a friend.
Your skills, technical or otherwise. This is what makes an idea yours rather than anyone's.
0 / 2000
Past employers, industries, internships, and the processes you watched waste everyone's time.
0 / 2000
Things that annoy you often enough that you would pay to make them go away.
0 / 2000
New technology, regulation or behaviour that makes something possible now that was not before.
0 / 2000
Fill in the first two fields: they are what founder-market fit is judged on.
What you write stays confidential and is only used to generate your ideas. You will be asked to create a free account to see them.
Where good startup ideas actually come from
The best ideas are rarely invented at a brainstorm. In the talk below, YC partner Jared Friedman lays out the "recipes" that produce them, and they all start from something you already have. This generator applies four of them, which is why it asks about you first.
Start from your unfair advantage
What can you or your team do that most people cannot? Ideas built on real expertise are ideas competitors cannot copy by simply deciding to.
Mine what you have lived
The jobs you have held, the industries you know, the problems you personally wish someone would fix. You already know things about them that outsiders cannot research their way into.
Ride a recent change
A new technology, a new regulation, a shift in behaviour. Something that recently became possible or necessary is an opening that did not exist for the incumbents.
The talk this tool is based on
His recipes became the questions we ask you. His evaluation criteria became the score on every idea you get back. Worth the 45 minutes.
How to Get and Evaluate Startup Ideas, by Jared Friedman (Y Combinator Startup School). Watch on YouTube
Every idea comes back evaluated against the questions YC actually asks, not just described. Here is what you get for each one:
The problem
Stated as a problem, never as a product looking for one.
Who has it
The specific customer, not a vague market.
Why you
Founder-market fit, tied to what you told us.
Why now
What recently changed to make this possible.
Market size
An honest estimate, not invented precision.
Problem intensity
How painful this is for the people who have it.
Competitors
Who is already there. Nobody there is a warning sign.
Tarpit warning
Flagged when the idea looks like one many teams have already died on.
Your first step
Something concrete you can do this week.
Already know which idea you want to pursue?
Turn it into an investor-ready business plan: market research, financial forecasts and a complete document, all in one place.
Build my full project →Frequently asked questions
This tool is the step before it. Use the generator when you want to start a company but do not yet know what to build; use the startup idea validator when you already have an idea and want to know whether it holds up. Most founders use them in that order: generate a shortlist here, then validate the one you like best.
A tarpit idea is one with a widespread problem and an apparently obvious solution that dozens of teams have already tried and failed at, for a structural reason that is not visible from the outside. Think of apps to coordinate plans with friends. They are dangerous precisely because they feel promising and attract smart founders for years. When a generated idea resembles one, we say so and name the obstacle rather than quietly leaving it in the list.
Because founder-market fit is one of the first things investors look for, and because an idea anyone could execute is an idea anyone will. Asking for a topic produces ideas you have no particular advantage in. Asking what you know produces a shorter list you are actually placed to win.
No, usually the opposite. Competitors prove someone is willing to pay to have the problem solved. An empty competitive landscape more often means there is no market than that you found an untouched one. What matters is whether you can be meaningfully different, which is why each idea comes with its existing players listed.
Your first generation is free with an account. Afterwards each run costs one token, as with the other AI tools in the app.
No. Your background is used only to generate your ideas. It is not published, sold, or used to train anything.