Researchers crack Google's 64-bit docid: it's a Jenkins hash fingerprint of the URL

gaganghotra_ · x · 2026-10-19

• An open-source project reverse-engineers Google Search's internal 64-bit docid, showing it's not an index counter but a fingerprint of the exact URL string.
• Formula: docid = (Jenkins32(url, seed=0) << 32) | Jenkins32(url, seed=102072), using Bob Jenkins' 1996 lookup2 hash in the exact variant of Google's internal gutil C++ library, whose code has been public for years in Apache Kudu (Apache 2.0 / 3-clause BSD).
• The hash input is the URL exactly as written — scheme, www, trailing slash, query string, case, and percent-encoding all matter.
• The same docid works across Search, Images, Discover & News; the hex form's first 10 characters match Google Images/Discover page ID prefixes.
• Includes formula, code, empirical validation and a reproducible Python implementation on GitHub; useful for SEO and search research.

Original post →

More from Research

• New evidence suggests consciousness is a dynamic flow, not a static thing — AnnaCiaunica · 2026-10-19
• MEAV: editing LLM alignment at inference time with composable alignment vectors — burkov · 2026-10-19
• Biomedical researcher hails GPT-5 science acceleration paper with OpenAI as his most consequential work — DeryaTR_ · 2026-10-19
• Frank Nielsen paper introduces scale-of-means framework unifying Fréchet and Karcher means — FrnkNlsn · 2026-10-19
• Dust hyper-optimization delivers ~10x speedup, verified; beyond-backprop algorithms coming — burny_tech · 2026-10-19
• Leading mathematicians warn: rushing to formalize all of math with AI could do lasting damage — AlexKontorovich · 2026-10-19

Research channel →