Sample counted 2026-09-01 off a 2,000-line synthetic fixture log, not a customer file.
What you getcounted 2026-09-01
You send us one of your own web-server log files (Apache or nginx Combined Log). We return a dated report that names:
- The share of requests that were named machines versus notOn the sample log: 725 named machines (36.25%) and 1,275 other (63.75%).
- Which named machines, by hitsA named machine means the request's User-Agent string contained a token from our existing 20-family list. It is not a claim about who the client really is. It is not Google Analytics.
- How many bytes those machines took, versus the restOn the sample log: 49,843,673 of 65,814,950 bytes (75.73%). Bytes are the size field in the log. We do not guess them.
- A delete-proof line for the file you sentWe delete your log within 24 hours.
Sample, counted off the 2,000-line fixture log
These numbers are from the pack's sample report. The log is synthetic. Invented addresses only. Not a live customer file. Parser verdict: MEASURED. Coverage on that fixture: 2026-08-25 to 2026-08-31.
| Fact | Number |
|---|---|
| Log lines | 2,000 |
| Named-machine requests | 725 (36.25%) |
| Other | 1,275 (63.75%) |
| Bytes taken by named machines | 49,843,673 of 65,814,950 (75.73%) |
| Rank | Named machine | Hits | Bytes |
|---|---|---|---|
| 1 | openai_gptbot | 180 | 15,308,142 |
| 2 | anthropic_claudebot | 120 | 8,645,871 |
| 3 | bytedance_bytespider | 90 | 8,194,614 |
| 4 | amazonbot | 70 | 4,482,625 |
| 5 | perplexitybot | 55 | 3,192,493 |
| 6 | commoncrawl_ccbot | 45 | 5,402,385 |
| 7 | meta_externalagent | 40 | 1,881,977 |
| 8 | openai_searchbot | 30 | 1,081,005 |
| 9 | openai_chatgpt_user | 25 | 551,525 |
| 10 | google_extended | 20 | 361,670 |
The other ten families on the list each had 5 hits (50 hits). Five Googlebot lines were in the log and were not named (not on the list). Named machines took about three-quarters of the bytes on about one-third of the requests.
What this cannot tell you
The upload page is not built yet. Today a person would take the log by email. The parser runs without a person. Delete-within-24-hours is our rule. The delete timer itself is not built yet.
It is not a claim about who the client really is. A named machine means the User-Agent contained a token from our existing 20-family list. Classic search crawlers that are not on that list (example: Googlebot) are not named. A rival with a longer list will look more complete.
We do not name the identity of any visitor.
The sample is not a customer log. It is a synthetic 2,000-line fixture. We did not use a real customer file to build it.
See the file
You do not have to take our word for what is in the report. Here is the top-10 named-machine table from the 2,000-line synthetic fixture, carrying the same columns as the sample CSV. Nothing in the counts is rounded.
- Open the sample as a CSVA plain spreadsheet file. Columns: machine_name, hits, bytes, share_pct. It saves to your machine rather than painting itself into a browser tab, and it opens in Excel, Numbers or Google Sheets.
- The same rows as JSONThe same rows again, laid out for reading with code.
Not on sale right now
Not on sale right now. We cannot yet deliver this automatically after payment, so we are not taking payment.
Email a questionNo payment is taken on this page.