Do not treat the file as a ranking requirement
Google’s generative-search guidance says its established SEO foundations remain relevant and that unnecessary AI text files such as llms.txt are not required. There is no basis for treating a missing file as a guaranteed visibility failure.
An optional machine-readable index can still be a convenient way to describe selected public documentation. That utility is different from evidence of search ranking or citation impact.
Start with evidence you can inspect
Check whether intended public pages respond successfully, whether crawler policy matches your intent, and whether indexing directives exclude the content. Review meaningful titles and page structure. Correct malformed structured data when it is present.
These checks describe technical readiness. They do not establish how an AI provider ranks or chooses sources.
If you publish an index
Choose public, stable documentation links and a concise description. Keep it current when those URLs change. Do not include private endpoints, credential-bearing links or claims that your site has not substantiated.
The llms.txt generator creates a local text draft. Review it before publication. It does not submit anything to an AI service.
Understand what the score excludes
AI Crawl Score assigns no penalty for missing llms.txt. It also avoids assigning a trust or citation-quality score merely because certain words appear on a page. See the scoring methodology for every weighted check.
If you investigate any effect of publishing the file, use a defined experiment and state its limits. A traffic increase after publication is not, on its own, proof that the file caused it.