Which AI systems read this site in the last 30 days, which of them did it to answer someone's question, where everyone connected from — and how many were people. People and machines are counted separately and never added together.
2026-09-16 – 2026-10-15 · updated 2026-10-15 19:36 UTC
Each row against its own peak — machines outnumber people many times over, and one scale would flatten the people row to nothing.
"Reads" — an AI company's crawler collecting pages to train a model or build the index its assistant searches. "Answers" — an assistant fetching a page at that moment because a person asked it something.
| Company | Headquarters | Reads | Answers |
|---|---|---|---|
| Amazon | US | 168,449 | 0 |
| Meta AI | US | 75,974 | 0 |
| ByteDance (Doubao) | CN | 23,805 | 0 |
| Perplexity | US | 22,106 | 0 |
| Claude | US | 20,395 | 283 |
| ChatGPT | US | 1,720 | 404 |
Page views by people who clicked a link to this site inside an AI assistant's answer.
| ChatGPT | 3 |
| Copilot | 3 |
From search engines, for comparison: 116
Reads by search-engine crawlers. Google's crawler also feeds Google's AI features.
| Yandex | 94,181 |
| Apple | 67,957 |
| Bing | 33,705 |
| 11,036 | |
| Naver | 79 |
| DuckDuckGo | 49 |
| Country | People | Views |
|---|---|---|
| US | 1,318 | 2,164 |
| BR | 1,185 | 1,186 |
| AR | 349 | 350 |
| GB | 326 | 490 |
| MX | 291 | 291 |
| BD | 286 | 289 |
| VN | 253 | 264 |
| FR | 239 | 556 |
| CN | 224 | 734 |
| TR | 221 | 221 |
| PK | 204 | 204 |
| CO | 191 | 192 |
| ES | 186 | 188 |
| ZA | 166 | 166 |
| RU | 164 | 166 |
The country of the data centre a crawler connects from — not where its company is (that is the Headquarters column above), and not where the person asking is. ByteDance, a Chinese company, crawls from Singapore, so it appears here as Singapore.
| US | 211,678 |
| IE | 75,950 |
| SG | 23,801 |
| GB | 1,633 |
| DE | 24 |
| CH | 16 |
| NL | 6 |
| JP | 5 |
| AU | 5 |
| IN | 4 |
| HK | 4 |
| KR | 4 |
| NZ | 3 |
| ES | 2 |
| BR | 1 |
| Language | People | Views |
|---|---|---|
| English | 6,667 | 9,101 |
| 한국어 | 2,395 | 2,520 |
A crawler is counted by the name it declares; requests that claimed a crawler's name and were found to be impersonating it are left out (50,401).
People are page views not marked as bots, after removing what is not a person:
| All non-bot page views recorded | 173,467 |
| − the site's own operator | 5,266 |
| − addresses whose first visit claimed to come from inside the site (a forged referrer) | 151,033 |
| − proxy swarms: one browser string, 100+ single-visit addresses in a day | 7,906 |
| = page views by confirmed people | 11,621 |
A public page on mygenelog.com that counts, for the last 30 days, which AI systems read MyGeneLog, which of them fetched a page to answer a person's question, which search engines indexed it, where people and crawlers connected from, and how many readers were people. People and machines are counted separately and never added together. It is rebuilt every half hour from the site's own server records.
A read is an AI company's crawler collecting a page with nobody waiting — to train a model or to build the index an assistant searches (ClaudeBot, GPTBot, Amazonbot, Meta-ExternalAgent). An answer is an assistant fetching a page at that moment because a person asked it something (ChatGPT-User, Claude-User, Perplexity-User). Answers are fewer, and they are the ones closest to a reader.
Crawlers are counted by the name they declare, and requests that use a crawler's name but are found to be impersonating it are left out. People are page views not marked as bots, minus the site's own operator, addresses whose first visit claimed to come from inside the site (a forged referrer), and proxy swarms — one browser string bringing a hundred or more single-visit addresses in a day. Each removal is shown in the "How this is counted" table.
The crawler-server table shows the country of the data centre a crawler connects from, not the country of the company. ByteDance, which is headquartered in China, runs its crawler from Singapore — nearly all of its reads come from Singaporean addresses — so it appears there. The company table's Headquarters column gives the company's own country.
Googlebot is Google's search crawler, and Google uses the same index for its AI features such as AI Overviews and Gemini. MyGeneLog counts Googlebot's reads as search indexing, and shows Google with the AI companies because that index is what its assistant answers from.
Every half hour. Each update covers the 30 days ending today and states when it was built, in UTC, under the figures at the top of the page.
Yes. They are published under CC BY 4.0. Name MyGeneLog and link to https://www.mygenelog.com/ai-visibility, with the date you read them — the figures move every half hour.