Is this a startup idea validator?

It is a decision tool, and the distinction is the point. A validator is built to assess the idea you bring; this is built to end the ones that will not work and to tell you what to do next about the one that might — including sourcing candidates itself when you do not have an idea yet.

Can it tell me my idea is bad?

Yes, and that is the intended outcome for most ideas. 12 deal-breaker checks run on every idea and any one of them can end it. The published sample run is worth reading precisely because of what it refuses.

There is no honest accuracy number, and anyone quoting one is inventing it. Nobody knows whether an unbuilt idea would have sold, so there is no ground truth to score against. What you can check is the reasoning: open the sample run and see whether the problems it cites are real and whether the verdict follows from them.

Will I get the same verdict twice?

Not necessarily. The ratings come from a model sampling at temperature 1, so they move between runs; the scoring arithmetic on top of them is fixed. Treat a borderline score as borderline rather than as a measurement.

What about business idea validation generally — not software?

The deal-breaker checks are about economics rather than code: whether a real problem exists, whether money repeats, whether one person can reach the buyer and support them. Those apply to a service or a physical product too. The worked examples on this site are software, so that is what the published runs show.

What a real startup idea check looks like

Most idea validators score your idea and agree with you. Here is what a check that can say no actually does, with a published run you can open and read.

What does a startup idea validator actually check?

A startup idea validator scores the idea you bring it. WhittleOS runs 12 deal-breaker checks where any single failure ends the idea, computes the score in code from seven rated areas instead of asking a model for a number, and binds the verdict to your time, money and skills. The published sample run shows what it refuses, not just what it approves.

Four differences that decide whether a check is worth running

It can return a no

A validator that scores every idea in the 70s has no decision in it — the number moves, the answer never does.

Every idea goes through 12 deal-breaker checks, and any one of them can end the run. A verdict that cannot be KILL is not a verdict.

The score is arithmetic, not an opinion

A language model is asked for a number out of 100, and it returns one that sounds right.

The model rates seven areas; the score is then computed in code from a fixed weighting, so the same ratings always produce the same total and grade. What the model cannot do is hand you a flattering number directly.

The problems come with somewhere to look

Market claims arrive as confident prose with nothing behind them.

The published run pulled 56 documented problems out of 41 sources, and 32 of them carry a link you can open. The rest are labelled as our estimate rather than dressed up as evidence.

The verdict is bound to your situation

The same idea gets the same answer regardless of who is asking.

Your time budget, capital, skills and support tolerance are inputs, so an idea that is workable for a funded team and unworkable for one person after hours does not get one shared answer.

The questions an idea has to survive

Every candidate is put through these. Any one of them can end it — that is what makes the answer a decision rather than a score.

Is there a real, documented problem behind it?
Is there a path to money that repeats?
Is there a repeatable way to reach the buyer?
Can it charge enough to be worth one person's time?
Can it run without hand-holding every customer?
Does it survive if one platform changes its rules?
Is there a product left if an AI vendor ships this feature?
Can you, specifically, sell and operate it?
Can it launch without licences or compliance audits?
Can it sell without legal review and a sales team?

For where the candidates come from in the first place, rather than how one is judged, see where startup ideas actually come from.

Read a run before you pay for one

A published single-idea check — The full verdict, the seven rated areas and the computed grade on one real idea.

A published Discovery run — 20 sub-markets, 24 pages read in full, and the problems that survived — with their sources.

Questions

Is this a startup idea validator?: It is a decision tool, and the distinction is the point. A validator is built to assess the idea you bring; this is built to end the ones that will not work and to tell you what to do next about the one that might — including sourcing candidates itself when you do not have an idea yet.
Can it tell me my idea is bad?: Yes, and that is the intended outcome for most ideas. 12 deal-breaker checks run on every idea and any one of them can end it. The published sample run is worth reading precisely because of what it refuses.
How accurate is it?: There is no honest accuracy number, and anyone quoting one is inventing it. Nobody knows whether an unbuilt idea would have sold, so there is no ground truth to score against. What you can check is the reasoning: open the sample run and see whether the problems it cites are real and whether the verdict follows from them.
Will I get the same verdict twice?: Not necessarily. The ratings come from a model sampling at temperature 1, so they move between runs; the scoring arithmetic on top of them is fixed. Treat a borderline score as borderline rather than as a measurement.
What about business idea validation generally — not software?: The deal-breaker checks are about economics rather than code: whether a real problem exists, whether money repeats, whether one person can reach the buyer and support them. Those apply to a service or a physical product too. The worked examples on this site are software, so that is what the published runs show.

The free founder read is the honest place to start: it reads your situation before any idea is scored.

Get a free founder read