Who runs this
C. Kerger built and runs the Fameproof measurement pipeline — the prompt panels, the weekly model runs, and the scoring that turns raw model output into a comparable index. Every audit that leaves this site is read and signed by a person before it is sent.
info@fameproof.comFameproof is deliberately small. One person designs the prompt panels, maintains the measurement pipeline, reviews every weekly run before it publishes, and writes the audits that are sold from this site. Where a number on this site is wrong, there is one person to hold responsible for it, and their name is on this page.
Why it exists
A growing share of the questions that used to go to a search engine now go to an assistant, and an assistant's answer names two to four options rather than listing ten links. Being inside that handful is starting to matter the way first-page ranking mattered — and almost nobody can tell you whether they are.
What passes for evidence today is a screenshot: one person, one prompt, one model, one day. Model output varies between runs, between versions and between phrasings, so a screenshot is closer to a coin flip than a finding. Fameproof exists to replace it with something repeatable: a fixed panel, declared in advance, run on a schedule, with the evidence published underneath the number.
A high score means an entity surfaced more often and earlier inside a defined test. It is not a claim about fame, quality or public approval, and the methodology page states plainly what it cannot tell you.
How it is measured
Each entity is asked about with a panel of 24 natural-language questions built for its category — never questions that name the entity, because a prompt that names you guarantees a mention and measures nothing. Each question is asked 4 times per model, and the entity's position in each answer is recorded.
The last completed run scored 293 entities on 5 August 2026, across ChatGPT and Gemini. The full panel behind any score is published on that entity's own page — for example Nvidia's prompt table shows all 24 questions and where each model placed it. Nothing on the public boards is derived, modelled or estimated.
The methodology page carries the full account: scope, panel design, scoring, ranking and the limitations that come with sampling a system that changes underneath you.
Paying does not move your score
Fameproof sells things, and none of them touch a published number. Buying an audit or a subscription changes nothing about where an entity sits on a public board: the panels are category-level and identical for every entity in that category, the runs are automated, and there is no mechanism — and no willingness — to move a published number for money. If that were on offer, none of the other numbers would be worth anything.
Corrections and removals
Everyone on these boards was listed without being asked. That comes with obligations, and they are not fine print:
Where to find us
Press and research enquiries reach the same inbox and are answered first — the underlying run data can be shared for reporting on request.
Common questions
Who runs Fameproof?
Fameproof is built and run by C. Kerger, founder & lead analyst, from 209 Hamilton St, San Francisco, CA 94134, United States. The same person builds the prompt panels, runs the weekly measurement and signs the audits.
Where do the scores come from?
Every score is measured, not estimated. A panel of 24 category questions is asked 4 times per model per entity, and the entity's position in each answer is recorded. The full panel for any tracked entity is published on its own page, prompt by prompt.
Can I have my listing corrected or removed?
Yes. Tell us which listing and what is wrong with it and we re-run the panel and correct it. Ask for removal and the page comes down — you do not need to give a reason.
Does Fameproof work for the AI companies it measures?
No. Fameproof is independent and is not affiliated with Anthropic, OpenAI or Google. It measures the public output of their products and has no commercial relationship with any of them.