llms.txt Observatory
A public, dated measurement layer for llms.txt. The figures describe a fixed panel and observable HTTP behavior, not all websites or AI usage.
Last updated:
Latest fixed-panel snapshot
51.8%
113 of 218 testable hosts served a qualifying file
14.0 KB
median file size across parsed directory files
79.6%
parsed directory files with a blockquote summary
The current observation was generated on . The panel is fixed and curated. It should be compared with future releases, not extrapolated to the web.
The observatory is designed for reproducibility. It records the public HTTP response, parse outcome and link checks available at release time, so a reader can distinguish adoption, syntax quality and discovery signals instead of treating a served file as proof of AI usage.
By sector
| Sector | Tested | Serving | Rate |
|---|---|---|---|
| AI & ML | 36 | 19 | 52.8% |
| Developer tools | 61 | 42 | 68.9% |
| Docs platforms | 7 | 1 | 14.3% |
| E-commerce | 13 | 7 | 53.8% |
| Fintech | 16 | 8 | 50% |
| Media | 14 | 0 | 0% |
| Other | 23 | 4 | 17.4% |
| SaaS | 48 | 32 | 66.7% |
V2 implementation signals
The v2 observation is narrower than a client-support matrix. It recorded 286 reachable links out of 294 checked, 31 Markdown
alternates, 4 describedby relations, and 21 hosts with a qualifying subpath
file.
Downloads and scope
Use the JSON release manifest for source hashes and limitations, the CSV for sector values, and the status endpoint for the current snapshot identifier. Read the full methodology before quoting a number.