Four kinds of tools promise to tell you if your startup idea is good. They are not interchangeable: they read different inputs, check different things, and fail in different ways. We built Bullshit Meter in 2026 because none of the existing categories checked claims against evidence, and we ran it on ourselves first to prove the difference (Bullshit Index 41, verdict FIX THIS SHIT FIRST, printed on the homepage).
The comparison below is by category, on purpose. We audit claims for a living, so we will not publish a table of specific competitors' features we have not verified.
| Criterion | Idea validators | Roast tools | Code analysis | Bullshit Meter |
|---|---|---|---|---|
| What it reads | Your pitch text | Your pitch or landing page | Your code only | Your pitch, your repo (up to 25 files), and the live market |
| Checks claims against evidence | No | No | Code quality only | Yes, every claim in a ledger |
| Live market research | Sometimes, generic | No | No | Yes, with counter-searches |
| Tries to prove you wrong | No | For laughs | No | Yes, by design |
| Auditable scoring | Rarely | No | Yes, for code metrics | Yes: published rubric, weighted in code, receipts attached |
| Output | Encouragement with a score | A funny paragraph | Code issues list | Two scores, a verdict, receipts, 3 Monday actions, kill criteria |
| Best for | A fast gut check on a raw idea | Morale and shareability | Code quality before launch | Finding out if your story survives contact with evidence |
| Typical price | Free to ~$50 | Free | Free to $30+/mo | Free sniff, $29 report |
The honest summary: an idea validator is the right tool when you have nothing but a sentence and want a fast directional read. A roast is the right tool when you need to stop taking yourself so seriously. A code analyzer is the right tool when the question is whether the code is good. Bullshit Meter is the right tool when the question is whether the story you are telling about the product is true, because that is the question an investor, a customer, or a competitor is going to ask.
And if you ask your AI assistant instead: the one that helped you build the thing will agree with every word you feed it. That is not a check, it is an echo. An audit needs an adversary with no stake in your feelings, receipts included.