A prompt-based business idea generator writes ideas from what you type, so every idea starts in the model, not in a problem anyone has written down. WhittleOS works the other way round. Its free Pain Receipts search returns problems our sweeps recorded from public pages, each labeled with where we found it, and a Discovery run turns a market into a short list of ideas with the rejects shown.
We search the problems the WhittleOS engine has already recorded from public pages. Type a niche, not your whole idea — and your text never goes in the URL.
Why a generated idea has nothing to check
An idea written by a model from a prompt describes a customer, a problem and a price that all came from the same place: the model. It reads well, because fluent text is what a model is good at. The trouble starts when you try to check it. There is no page to open where the problem is described, no buyer to point to, and nothing that separates a real problem from a plausible one.
That is fine for brainstorming, and a generator is quick at it. It is a poor basis for months of building, because the first real question — does anyone have this problem badly enough to pay for a fix — is the one a generated idea skips.
What the search on this page returns
Type a market or a job — freelance invoicing, wedding photographers — and Pain Receipts looks through the problems our sweeps have recorded from public pages: 1,823 of them on 2026-09-29. Each result is labeled "from a real page" when we read the whole page, or "from search" when we saw the search result but not the page, and links to the page whenever we captured one.
A result is a problem, not an idea. It tells you that a public page describes the problem, and which page: sometimes a forum post or a review, sometimes a product listing where the gap is visible rather than stated. That is the part a generator cannot give you. If we have not recorded your market yet, the search says so, or shows the closest rows it has and marks them as a partial match. It never makes a result up.
To see what that looks like at scale, the problems page lists one recurring problem for each market we track. On 2026-09-29 the largest count among them was collections, dunning or late payment in invoicing & payments: 41 of the 347 problems recorded there carried that theme.
From problems to ideas, with the rejects shown
A problem is not a business yet, and turning one into the other is most of the work. A Discovery run does the first part of it for a market you name: it plans a set of related markets around yours, searches and reads public pages, drafts up to 150 candidates and keeps at most 5. The run page shows why candidates were dropped, and the first check on its list asks: "Is there a real, documented problem behind it?"
What comes back is a short list scored against your founder profile, each idea with a verdict and a way to test demand without a sales call. One market costs 2 credits from a paid pack — the free signup credits do not cover a Discovery run — and a complete run is published on this site, unedited, if you want to read one before paying anything.
When a generator is the right tool
If you want a pile of ideas to react to in the next minute, a generator is faster, and nothing here competes with that. Use it for range.
Then take the idea you like best and type its market into the search above. If it finds the problem, you have a page to read before the first line of code. If it finds nothing, that means we have not recorded that market yet, not that nobody has the problem.
What this cannot tell you
Whether an idea will sell. A recorded problem shows that a public page describes it; whether enough people will pay to fix it is a test, not a search. The search and the problems page give you the first half. Asking for money before you build is the second half, and no tool, ours included, can do that part for you.
Common questions
Is there a free business idea generator?
The search on this page is free and needs no signup, but it does not write ideas — it shows the problems ideas can be built on, with their sources. If you want directions suggested for you, the founder read looks at your profile and names the kinds of business that fit it. It costs 1 credit, covered by the 2 free credits every new account gets.
Can AI generate a good business idea?
It can write a plausible one in seconds. Whether it is good depends on things the prompt never saw: whether the problem is real, whether you can reach the buyer without a sales team, and whether the price covers your time. Those are checks, and each one needs evidence from outside the model.
What if my market is not in the search yet?
Then it says so, or shows the closest rows it has and marks them as a partial match. The search only shows problems our sweeps have recorded, so an empty result means we have not recorded that market, not that the problem does not exist. A Discovery run on that market searches it, and the problems it records join the same library.
See it rather than take our word
Everything this page describes is published somewhere on this site, unedited and without a signup — a complete run, a full report, and the checks that produced them.