AI bot traffic on the sites we operate
This page counts requests whose user agent named a known AI or search bot, on the websites we operate ourselves (a handful of mostly small and new sites). It shows counts per bot and page type for the last 30 days and nothing else: no IP addresses, no user-agent strings, no full URLs.
The data is still thin. 290 counted requests on 1 day since 2026-10-19, from 4 sites. That is too little to compare bots or to see trends; read the numbers as a sample, not as a measurement of the web.
Requests by bot
| Bot | Owner | Requests | Share |
|---|---|---|---|
| GPTBot | OpenAI | 158 | 54 % |
| ClaudeBot | Anthropic | 70 | 24 % |
| Bingbot | Microsoft | 28 | 10 % |
| Googlebot | 26 | 9 % | |
| OAI-SearchBot | OpenAI | 5 | 2 % |
| Claude-User | Anthropic | 3 | 1 % |
What they request
| Page type | Requests |
|---|---|
| Content pages | 191 |
| robots.txt | 27 |
| Sitemaps | 22 |
| Legal pages | 17 |
| Home page | 15 |
| Images and files | 12 |
| RSS feeds | 4 |
| llms.txt | 2 |
Requests per day
| Day (UTC) | Requests | Bar |
|---|---|---|
| 2026-10-19 | 290 |
Open data
The list of bots on this site (names, owners, tokens, sources, check dates) is open data under CC BY 4.0: JSON or CSV; see the column guide and how to cite it.
How to read this
- A request is counted when its user-agent header contains the documented token of a bot we track. A user agent is a claim: anyone can send it, so some counted requests may not come from the operator.
- We count requests, not visits and not people. One crawl can mean many requests.
- The sites are few, small and new, so the numbers say little about the web in general.
- Only counts per day, site, bot and page type are stored. See the methodology and the privacy policy.
Which of these bots can read your site? Run the checker or browse all bots.