01
Basic
TSP Basic — Verifiable AI evidence
Why ordinary logs are not proof, how the Trust Standard Protocol makes AI evidence durable and verifiable — and what TSP explicitly does not promise.
The introductory track to the Trust Standard Protocol (TSP) — an open protocol for verifiable AI evidence. You will learn why ordinary logs do not hold up as proof, how a TrustEnvelope is sealed, anchored and verified, and why the cryptography (Ed25519 signatures, SHA-256 over canonical JSON and a hash chain) makes the evidence tamper-evident and verifiable offline by anyone. Just as important: you learn the limits — TSP reports evidence integrity, not truth, legality or compliance. TSP is in v1.0 public preview, and the track is honest about what is finished and what is pilot.
Content
Why TSP
The "logs are not proof" problem, the three futures for AI evidence and the regulatory backdrop.
-
Logs are not proof — the problem with ordinary logs7 min
-
Three futures and the regulatory backdrop6 min
-
Quiz: Why TSP5 min
How it works
The TrustEnvelope flow sealed → anchored → verified, and the cryptography explained for non-engineers.
-
The TrustEnvelope in three steps7 min
-
The cryptography for non-engineers — and "Break one byte"7 min
-
Quiz: How it works5 min
The limits
The doesNotMean list, "verification is not truth", self-attested vs independently attested — and what the certificate documents.
-
What TSP does not mean — the doesNotMean list7 min
-
Self-attested vs independently attested6 min
-
Certificate basis5 min
-
Reflection5 min
-
Quiz: The limits5 min