For the complete documentation index, see llms.txt. Every page on this site is also served as Markdown: append `.md` to any URL, or send `Accept: text/markdown`.
Skip to content
Attensira LogoAttensira
Book a demo

See which pages decide the answer, and get onto them

Citation tracking records every link an AI answer attaches: how often engines link your own pages (your citation rate), which domains they read instead, and which sources decide each question you track, ranked by weight, each with its next move: fix your page, get listed, reply in the thread or pitch.

Start freeBook a demo

7 days free · card required

Deciding sources · “AI crawler logs analysis tool”attensira.com · not named in this prompt yet · n = 105
ChatGPT · n = 57 · not on any of the top three
github.com/Cairrot-Inc/cloudflare-ai-crawler-tracker
New
developers.cloudflare.com · analyze AI traffic
New
aboutmartech.com · best AI crawler analytics tools
New
Perplexity · n = 48 · move: be in the roundup
searchengineland.com · log file analysis for AI crawlers
New
edgecomet.com · log analyzer
New
aisearch.similarweb.com · log file analysis
New

Shown as a baseline, not a result.

Analyst agentAnalystattensira.com workspace, read over the Attensira MCP · · 30 days to 2026-10-04, 24 of 31 days sampled

  • Analyst agentAnalyst · Ranks the sources behind each answer by weight, and marks the ones you are already on.

  • Outreach agentOutreach · Takes a source you are not on and plans the placement, from your own mailbox.

How does citation tracking work?

Mentioned is not cited: an engine can name a brand it never links. Citation tracking reads the links, then turns them into a list of places to be.

  1. Collect every link

    Footnotes, source lists, inline references and source panels are normalised to the same thing: a URL, attributed to a domain. Each answer is stored with its links.

  2. Count your citation rate

    Answers that linked a page on your domain, divided by every answer read, the same denominator as mention rate so the two compare directly. Four links in one answer count once.

  3. Rank what decided it

    For each prompt and engine, sources are ranked by citation share × position, not raw counts: a page linked first in every answer decided more than one linked twice near the bottom.

  4. Name the next move

    You are not on it: new. A rival ranks above you: better. You are on it and ahead: fix your own page. Each move opens a chat where the crew plans and does the work.

Which domains do the engines link most?

Across every answer we track, a handful of help centres and vendor blogs carry the most links. Our own domain is seventh, behind two OpenAI domains, with about 1.5% of all citations. The chart is the top of the source board, ranked by citations, before any placement work.

Citations by domainn=105,376 citations
ChatGPT and Perplexity, 61 prompts · to
NameCitations by domain
help.openai.com5.3%
developers.google.com4.4%
semrush.com2.3%
otterly.ai2.2%
peec.ai2.1%
openai.com1.5%
attensira.comyou1.5%
Analyst agentAnalyst · Ranked the sourcesattensira.com sources board, read over the Attensira MCP · · n=105,376 citations · 25 of 31 days sampled

What kind of source decides on each engine?

Engines lean on different kinds of page. Until your own workspace has 20 reads of a source, the move comes from this table, and it says so. You count as on a source only on your own domain, or once the crew has verified a placement is live.

What kind of source decides on each engine?
EngineThe source it leans onThe move
ChatGPTYour own docs and pricing pagesFix your own page
ClaudeReview sites and listiclesGet listed
Perplexity, Gemini, Google AI OverviewsThreads, YouTube and roundupsBe in the thread or the roundup

What does citation tracking not do?

It reports links the engines attached. It does not guess at links they might have read but did not show.

  • A model naming you from memory has no link to attach, so citation rate is almost always lower than mention rate. The gap is the signal, not a bug.
  • Country is not part of the sources map: a page that decides a question in the US decides it in the UK too.
  • A source's type is inferred from its domain and cannot be corrected yet.
  • Pressing a next move starts nothing on its own. It opens a chat with the move written in, and nothing is sent until you press send.

Which pages did one answer about us link?

Ten links on one Perplexity answer: six to attensira.com and its docs, three to the help centre of Attensi, a different company, and one to tga.gov.au. Each link is counted against the domain it points to, so a look-alike source shows as its own row.

Answer · Perplexityasked

Question: what is Attensira

Perplexity answered: Attensira is an “AI search visibility” platform/agent that checks whether AI assistants name and cite your brand, then generates a content fix (as a pull request or draft) for you to review and approve.

Sources cited

  1. 1attensira.com (Attensira)
  2. 2attensira.com/about (Attensira)
  3. 3docs.attensira.com (Attensira)
  4. 4tga.gov.au
  5. 5help.attensi.com
  6. 6help.attensi.com/…/Intro-to-Attensi-Products
  7. +4 more

NamedAttensira is in this Perplexity answer.

Analyst agentAnalyst · Counted the linksattensira.com stored answer, read over the Attensira MCP · · 10 links, in the order shown

What else do people ask about citation tracking?

What is the difference between a mention and a citation?
A mention is your brand named in the answer text. A citation is a link the engine attached to a page on your domain. One answer can do either, both or neither, and each is reported as its own rate with its sample size.
My mention rate is high but citations are low. What does that mean?
You are known but not read: engines describe you through other people's pages. The lever is publishing pages that answer the tracked questions directly, and getting onto the sources that do decide them.
How do I get cited by ChatGPT?
Start with the sources map for the prompts you lose on ChatGPT. Your own docs and pricing pages decide most of its answers, so the usual first move is fixing your page so it answers the question; the crew can open that fix as a pull request or CMS draft.
Which of my pages are cited most?
The Pages board lists your own pages by how often they were linked, with the prompts that cited each one and when it was last cited and last crawled.
Can I pull the sources into my own tools?
Yes. The deciding_sources MCP tool and GET /v1/deciding-sources return the same map, narrowed to one prompt and window if you want. Reading it spends no credits.

See who the engines read instead of you.

Start freeBook a demo

7 days free · card required