About CiteFiles
A growing share of the questions people used to type into a search box are now asked of an assistant. The answer gets written from whatever that assistant could reach and read. If your site was not reachable, or was reachable but unreadable, you are not in the answer — and unlike a search ranking, there is no position twelve to console yourself with.
The awkward part is that this fails silently. Your site looks perfect in a browser. Meanwhile your CDN may be returning a bot challenge to ClaudeBot, or your platform may be answering 200 OK with an HTML page for /llms.txt, so every checker that only looks at the status code tells you the file is there.
So we ask twice — once as a browser, once as each crawler — and compare. Then we hand a model only the text a crawler would get and ask whether it can answer basic questions about you. Then we write the files you are missing, from your own words, and hand them straight back.
What we will not do
We do not invent facts about your business. Every line of a generated file traces to text your site already publishes. Where we have no source — your social profiles, your licence, your correction policy — you get a visible TODO, not a confident guess. A file full of plausible fabrications is worse than no file, because you would publish it.
We do not bypass anything. We identify ourselves honestly, request only public pages, and read one site at a time. In the crawler test we send real crawler user-agent strings, which is the only way to observe how an origin treats them, and we request the same public homepage a browser gets. We do not solve challenges, we do not rotate addresses, and we do not touch anything behind a login.
We do not tell you to open the gates. Blocking AI crawlers is a legitimate position, and for some publishers it is the right one. We will tell you what your current settings actually do, which is often not what their owner believes, and leave the decision where it belongs.
Who is behind it
CiteFiles is a small independent project, not a venture-backed platform with a sales team. It runs on our own hardware, including the language models — nothing you scan is sent to a third-party model provider. Questions, corrections and complaints all go to the same place: hello@citefiles.com, and a person reads it.
Try it on your own site
Free, and the summary needs no account. Read the rubric first if you would rather know how it is judged before it judges you.