Citelane

Hi Fazier! I built Citelane after checking my own small app: I asked four AI assistants the questions my buyers ask, 24 answers in total, and my app was mentioned zero times. The assistants cited a handful of information sites instead. Most AI visibility tools are dashboards. Citelane focuses on the step after: it shows which pages get cited instead of you, checks whether AI crawlers can read your page, drafts one page edit using only facts you confirm, and then checks that the edit is actually live. The check is free and needs no sign-in. I'd love feedback: which buyer question would you test first?

Good product, helped me alot

Does the report distinguish the API model tested from the consumer ChatGPT or Claude app? Their answers can differ, so I would want that distinction visible when using the report as a discovery baseline.

I like that Citelane reports repeated answers with uncertainty ranges instead of a single snapshot, since AI answers vary a lot between runs. Could you share how many repeats are typically used per question in the free check? One suggestion: showing a short plain-language note beside each range would help non-technical founders interpret it.

Reporting 95% uncertainty ranges on repeated answers is the part most AI visibility tools skip: one assistant answer is an anecdote, repeated runs are a measurement. A question on method: how many repeats do you run per buyer question by default, and do you vary the wording of the question as well, or only resample the same prompt? Small wording changes can change who gets named more than resampling does, so I would be curious which of the two your ranges cover.

Congrats on the launch! I like that the free AI crawler checker covers GPTBot, OAI-SearchBot, PerplexityBot and Claude with no sign-in, since that's an easy first win for founders. One suggestion: could the shareable report include a short "top 3 fixes" summary, so non-technical users know what to change first?

Repeated answers with 95% uncertainty ranges is a smart choice, since a single AI answer is close to random. How many runs per question do you do by default? Can I re-run the same question set weekly to see whether a page edit actually changed citations?

The detail I find most interesting is the approval tied to a hash of the exact drafted text, plus the check that the approved wording actually shipped on the live page. Most tools stop at reporting. One question: when you show 95% uncertainty ranges across repeated answers, how many repeats per question do you run by default, and does the free check use the same number?

The distinction between being mentioned and being cited is useful. I’d put the exact cited page and supporting passage beside each competitor, so a founder can see which claim won and replace it with a clearer, sourced answer.

The claim check that flags unsupported numbers or promises before a page edit ships is a smart guardrail. How many repeated runs per question do you use to get the 95% uncertainty ranges, since AI answers can vary a lot between runs?

Ran it on my studio's site: every technical check passed, yet 0 of 9 answers mentioned us — the assistants went to Upwork and Codeable. That split between "AI can read you" and "AI recommends you" is exactly what most audits miss.

The claim check that flags unsupported numbers or promises before a page edit ships is a smart guardrail against AI-written fluff. After it verifies the live page, can it re-check later in case a CMS update or rollback quietly changes the approved text?

Showing repeated answers with uncertainty ranges is the right call: a single "not mentioned" run from an assistant is mostly noise. One question about the crawler checker: does it look past robots.txt at CDN-level bot blocking? Cloudflare's AI crawler setting can turn GPTBot away even when robots.txt allows it, and that is easy to miss when you only read the file.

Running the same question several times and showing a range makes sense, because asking an assistant once can tell you almost anything. How many runs per question does it take before the range gets tight enough to trust? And do you see big differences between Perplexity and Gemini for the same page?