How these records are anchored

This site's central claim is that its verdicts were written down before the outcomes were known. Every page so far asks you to take our word for that ordering: our database says the verdict came first, and our database is ours. Nothing stops anybody writing yesterday's date into today's row.

So the ordering is taken out of our hands. Each night a single fingerprint is computed over the whole evidence ledger and submitted to public timestamping calendars, which fold thousands of such fingerprints into one Bitcoin transaction. After that, the statement "these exact records existed no later than this block" is checkable by anyone, with open tools, and does not depend on trusting this site at all.

What an anchor proves, and what it does not

It proves existence and ordering. These records existed by that time, in exactly this form. Change one byte of one row and the fingerprint no longer matches.

It does not prove the records are correct. A verdict can be anchored and still be wrong. What anchoring removes is the one thing a database can never establish about itself: that it is not being rewritten after the fact. A copy of a database proves nothing about when it was made.

The anchors

AnchoredRecords coveredFingerprint Calendar
01 Sep 2026, 08:50 UTC 5315 326e25c1094f925a…52313378 https://a.pool.opentimestamps.org
01 Sep 2026, 08:47 UTC 5315 e79a2bec230dcf57…22166e7d https://a.pool.opentimestamps.org
01 Sep 2026, 08:44 UTC 5298 b9a6422358f47099…7d0afb57 https://a.pool.opentimestamps.org
01 Sep 2026, 08:05 UTC 5268 ccf2bf4fbe77e185…97e838fc https://a.pool.opentimestamps.org
01 Sep 2026, 02:31 UTC 5163 1d90c553478064f4…71db6e6f https://a.pool.opentimestamps.org
31 Aug 2026, 02:31 UTC 5070 15f5269db41f095f…47fc7e41 https://a.pool.opentimestamps.org
30 Aug 2026, 12:20 UTC 5067 828345fe72be84a0…e3cd22dd https://a.pool.opentimestamps.org
30 Aug 2026, 12:05 UTC 5067 191a435ba9405723…5c4ea02f https://a.pool.opentimestamps.org
30 Aug 2026, 11:56 UTC 5067 a6f2ed38fe4932bd…9886be7d https://a.pool.opentimestamps.org
30 Aug 2026, 11:53 UTC 5067 666b66646c03312c…49fac6dc https://a.pool.opentimestamps.org
30 Aug 2026, 11:50 UTC 5048 a0ee81f49e058bda…3156a191 https://a.pool.opentimestamps.org
30 Aug 2026, 10:41 UTC 5034 07c5862b5babf608…0c1dc761 https://a.pool.opentimestamps.org
30 Aug 2026, 02:30 UTC 4954 03df007619713e76…dea85d45 https://a.pool.opentimestamps.org
29 Aug 2026, 10:59 UTC 4952 8316c8ad09f15bc3…a81f0edb https://a.pool.opentimestamps.org
29 Aug 2026, 09:53 UTC 4940 a853934ea9362eeb…c682be9f https://a.pool.opentimestamps.org

Verifying one yourself

The fingerprint is a Merkle root over every row in the evidence runs, verdict and outcome tables, hashed field by field in a fixed order. Rebuild it from the open dataset at /data, then check the timestamp proof with the OpenTimestamps client:

pip install opentimestamps-client
ots verify vetmytoken-anchor-<id>.ots

The client speaks to the same public calendars and to a Bitcoin node; it never asks this site anything, which is the point.

Anchoring costs nothing and requires no wallet: the calendars aggregate many digests into a single transaction. That is why it can run every night rather than when a budget allows.