Adding llms.txt did not increase AI citations: how does it relate to search-index eligibility?

Identify Google versus other intended consuming systems

This applies when the file is published and accessible without search changes. Identify the intended product and its documented reading behavior. AI services and coding agents can discover content differently. One tool reading the file does not imply all AI search uses it, and a logged visit is not a citation.

Separate documentation files from crawl and index controls

Check crawling and noindex, then canonical and preview controls against basic eligibility. Inspect llms.txt links separately from indexing. It replaces neither robots.txt nor access controls. Instructions not to read a location do not establish protection of private material.

Prioritize indexable content and useful evidence

For Google, repair actual indexing and content problems and provide verifiable answers, sources and useful internal links. Maintain the file for other consumers where justified, with documented purpose and working links, without promised ranking gains. Do not publish private Markdown, accounts or unauthorized material merely to increase claimed AI readability.

Assess changes with actual indexing and impression records

Check content eligibility with URL Inspection and observe comparable impressions and visits while logging file access separately. Record product, time and questions for manual observations rather than attribute an incidental citation to the file. A working or crawled llms.txt and a longer directory do not prove improved Google AI visibility.

This article created no file or ran an impression experiment and does not guarantee other systems follow one proposal. Current Google guidance says maintaining llms.txt neither helps nor harms its visibility or rankings; other consumers can differ. Keep checking dates as proposals and platforms change. The file is not access control or a substitute for content and verified outcomes.