Not sure where to start? Take a guided tour.
AI Site Grader
AI Query Scanner
llms.txt Generator83.7% of connected sites saw OAI-SearchBot in the last four weeks.
OAI-SearchBot reached 87 of 104 sites connected to Trakkr in the four weeks to 25 October 2026. This describes Trakkr’s sample, not the whole web.
Which AI crawlers reach the most sites?
Share of connected sites with at least one request from each crawler in the last four weeks.
27 published rows. As of 25 Oct 2026. The bars use one scale from zero; the JSON and CSV keep the source precision.
| Crawler | Operator | Purpose | Sites reached, last 4 weeks | Sites, last 4 weeks | Request share, 13 weeks | Requests, 13 weeks |
|---|---|---|---|---|---|---|
| OAI-SearchBot | OpenAI | AI search crawler | 87 | 10.4% | 3,455,251 | |
| ClaudeBot | Anthropic | Training crawler | 87 | 8.4% | 2,787,760 | |
| GPTBot | OpenAI | Training crawler | 70 | 6.5% | 2,143,347 | |
| Amazonbot | Amazon | Other crawler | 63 | 31.5% | 10,439,312 | |
| PerplexityBot | Perplexity | AI search crawler | 58 | 6.8% | 2,248,664 | |
| ChatGPT-User | OpenAI | Live fetcher | 57 | 6.1% | 2,022,620 | |
| Applebot | Apple | AI search crawler | 52 | 7.5% | 2,469,949 | |
| Bytespider | ByteDance | Training crawler | 50 | 8.6% | 2,835,910 | |
| YouBot | You.com | AI search crawler | 40 | 4.9% | 1,610,634 | |
| Claude-Code | Anthropic | Live fetcher | 40 | 0.6% | 194,155 | |
| cohere-ai | Cohere | Live fetcher | 39 | 0.2% | 53,017 | |
| GrokBot | xAI | Training crawler | 38 | 0.4% | 123,247 | |
| DeepSeekBot | DeepSeek | Training crawler | 38 | 0.1% | 44,081 | |
| CCBot | Common Crawl | Training crawler | 37 | 0.3% | 93,279 | |
| Perplexity-User | Perplexity | Live fetcher | 34 | 0.3% | 89,161 | |
| Scrapy | Zyte | Other crawler | 34 | 0.1% | 34,921 | |
| Meta-ExternalFetcher | Meta | Live fetcher | 31 | 0.1% | 17,261 | |
| Claude-User | Anthropic | Live fetcher | 28 | 0.3% | 103,476 | |
| OAI-AdsBot | OpenAI | Other crawler | 28 | 0.1% | 43,893 | |
| MistralAI-User | Mistral AI | Live fetcher | 28 | 0.1% | 28,686 | |
| Claude-SearchBot | Anthropic | AI search crawler | 20 | 1.2% | 387,542 | |
| Meta-ExternalAgent | Meta | Training crawler | 19 | 5.4% | 1,782,555 | |
| Google-Agent | Live fetcher | 19 | 0.0% | 8,744 | ||
| Diffbot | Diffbot | Other crawler | 13 | 0.2% | 82,320 | |
| Timpibot | Timpi | AI search crawler | 12 | 0.0% | 5,034 | |
| anthropic-ai | Anthropic | Training crawler | 7 | 0.0% | 13,715 | |
| Claude-Web | Anthropic | Live fetcher | 3 | 0.0% | 1,626 |
- Sites, last 4 weeks
- 104
- Sites, 13 weeks
- 134
- Requests from published crawlers
- 33,120,160
- Measurement window
- 27 Jul to 25 Oct 2026
Every number needs a label.
Read the sample, denominator and collection window before using this benchmark.
How this benchmark is built
Site reach is the share of connected sites that recorded at least one request from the crawler in the last four complete weeks (28 Sep to 25 Oct 2026, 104 sites). Crawlers are ranked by it.
Request share is each crawler's part of all requests from the 27 published crawlers over 13 complete weeks (27 Jul to 25 Oct 2026, 134 sites). A few busy sites can dominate it, so read it beside reach.
Finalized daily crawler summaries from sites connected to Trakkr, grouped by the crawler signature detected in each request. Trakkr’s own site is excluded. Only finalized days are used.
Purpose labels come from each operator’s own documentation, through the Trakkr bot registry.
This is a sample of sites connected to Trakkr, not the whole web.
A matching user-agent does not prove the named operator sent the request.
Counts are classified requests, not unique pages, training use or citations.
Crawlers seen on fewer than 10 sites are not shown, and weekly medians need at least 5 sites.
Classic search crawlers (Googlebot, Bingbot) and generic product labels are left out, because they are not collected the same way on every site.
The number of connected sites changes from week to week (50 to 98 sites in this window).
Earlier Trakkr crawler figures, such as Study 003 (June 2025 to February 2026), used a different sample and measure. Do not compare them directly.
- Snapshot updated
- 25 October 2026
- Generated
- 26 October 2026
- Published rows
- 27
- Licence
- CC BY 4.0
Data sources
Open dataset under CC BY 4.0. Keep the definition and source date with any quoted figure.
Crawler signatures are observations in this sample. They do not verify operator identity or establish a later citation or visit. The same weekly data drives the crawler profiles.
Questions, answered.
Definitions and publication limits for this dataset.
A single busy site can send millions of requests from one crawler and swing request share. Site reach counts each site once, so it shows how widely a crawler visits. The table keeps 13-week request share beside it.
The Agent answers from our docs, methods and pricing, and links to the source.
NikeAI crawler requestsOut of every 100 crawler requests
- GPTBot19,76641%
- OAI-SearchBot13,01727%
- ClaudeBot9,16019%
- PerplexityBot6,26713%
Watch the bots that read your site.
Trakkr logs the AI crawler requests your pages get, then checks citations and AI referrals separately.
- 14-day free trial
- Cancel anytime
- Works with Google Analytics