Timestamping a Giant Record of the Web

Project Timestamper has cryptographically timestamped all 339 billion web page snapshots in the Common Crawl collection to the Bitcoin blockchain, proving each snapshot existed byte for byte as of 2026-09-26. The project hashed roughly 134 million Common Crawl index blocks over a few weeks and attested those hashes in a single batch via the OpenTimestamps service, so verifying any one of the 339 billion snapshots takes a few seconds and a few hundred KB of downloaded data. The work aims to make later alterations detectable as AI-generated content and database tampering grow harder to distinguish from original web captures.

Our latest announcement from Project Timestamper concerns a number whose scale far exceeds anything we have worked with before. That number is 339 billion, or 3.39×10<sup 11</sup . It happens to be the number of web page snapshots captured by the Common Crawl https://commoncrawl.org/ project over the past 18 years. This epic web crawl is one of the most important in the world, perhaps second only in size to the continuous crawl that is performed by Internet Archive's Wayback Machine https://web.archive.org/ . Common Crawl's hundreds of billions of snapshots save web pages precisely as they existed over the years, totalling more than 10 petabytes of snapshot data. One might think of it as a recording of a significant part of the collective thoughts of humanity over the past two decades. Such a recording is obviously precious to us humans. It would be a shame if anything happened to it As AI's capability to generate plausible data and hack databases continues to grow, how will we know what's original and what is modification, imitation, confabulation? So we're happy to announce that we have cryptographically timestamped the whole thing Every web page capture has been committed via a chain of hashes to the Bitcoin blockchain, so there is proof that each captured page snapshot whenever it was captured existed byte for byte in the Common Crawl collection as of 2026-09-26. Any alteration after that date will be detectable. Ordinarily, timestamping the entire collection of individual snapshots would have been prohibitively slow: there are far too many items to process in a reasonable time. Fortunately, Common Crawl had already hashed all web page snapshots and collected the hashes in blocks of up to 3000 index entries. We were thus able to hash each of about 134 million blocks in just a few weeks, and timestamp those hashes in a single batch using the OpenTimestamps https://opentimestamps.org/ service. Our resulting hashes and timestamp attestations are arranged in such a way that cryptographically verifying any one of these 339 billion snapshots requires just a few seconds and a few hundred KB of downloaded data. Our verification tool https://github.com/project-timestamper/stamper lets you check for a snapshot's existence in a single command: npx tsx verify.ts --collection common crawl blocks --capture 0 https://en.wikipedia.org with a quick response indicating success or failure: Success Capture 0 https://en.wikipedia.org/ at 20260915155254 is in CC-MAIN-2026-39, its CDX block is attested by Bitcoin block 968666 000000000000000000010fb95a4fa547171edc525077191cbb74dc54acfaf0aa as of 2026-09-26T09:35:20Z, and the WARC payload SHA-1 matches 3I42H3S6NNFQ2MSVX7XZKYAYSCX5QBYJ We're grateful to the Common Crawl Foundation for their work building and maintaining the web capture collection, and for providing open APIs and open data. We hope to keep timestamping new data as the Common Crawl continues in the coming months and years.