Google's Crawl Stats beside your log

Google's Crawl Stats report is the only place Google says how much it crawled, and it has no API. Consultants export it by hand on report day and reconcile it in a spreadsheet; that is where a 64% "Other file type" headline turned out to be mostly redirects, and where a forum host Google crawled every day turned up that the log had never seen.

What changed

• On Search Performance, under the index-vs-fetch reconciliation, drop in the CSV files from the Crawl Stats export (Settings → Crawl stats → Export). Name the host for a domain property; one import per host.
• Google's crawl requests, download size and response time sit beside the log's verified Google fetches for the same Pacific days (Google's days), with the share the feed saw.
• The response, file type and Googlebot type tables show Google's share beside the log's, with the gap in points.
• Hosts Google crawls that the feed never sees are flagged. So are hosts the log sees that Google's table does not list.
• Two caveats travel with every result: Google's "Other file type" includes redirects, and fetches Google counts while robots.txt was unavailable never reach a log.
• API: POST and GET /public/v1/websites/{id}/seo/crawl-stats. MCP 1.30.0: import_crawl_stats, get_crawl_stats. CLI 2.10.0: seo crawl-stats-import, seo crawl-stats.