Explore (579)Live Systems (52)Pricing
Log InGet API Key✓ Verify It Yourself
H33  ·  verify-the-story  ·  playground  ·  v0.4
The evidence economy.

Verify a story in your browser.
No server. No daemon. No trust required.

Identity. Monitoring. Interdiction. Evidence. Examination. The first three are crowded — Chainalysis, TRM, Elliptic, Sardine, Merkle Science. The last two are the next market. This page is the evidence and examination layer running entirely in your browser.

Drop a replay bundle or a signed verification report. Both get checked locally — the bundle by the same Rust verifier the CLI runs (compiled to WebAssembly), the signed report by a standards-conformant ML-DSA-65 signature check in pure JS. Zero network calls after this page loads.

Trust boundary: Replay verification runs in WebAssembly — the exact same Rust core that drives h33-replay-verify. Signed-transcript verification runs in JavaScript against the published ML-DSA public key embedded in each transcript; any FIPS-204-conformant verifier produces the same answer.
Drop a replay bundle or signed report JSON file here
…or click to choose. Or paste JSON below.

What this proves

Replay bundle (WASM)

Same 10-check protocol as h33-replay-verify. Schema parses, timeline orders, every chain hash recomputes, frame refs resolve, continuity is consistent, no orphans, hash algorithms recognized, no cross-tenant rows, substrate bindings recompute. Bundle is bit-identical to what H33 recorded at export time.

Signed report (JS)

ML-DSA-65 (FIPS 204) detached signature over the embedded canonical payload, verified locally against the public key shipped inside the transcript. Confirms a specific verifier identity (by SHA3-256 fingerprint) ran vverifier_version against a specific bundle hash at a specific time and produced exactly the stated per-check verdicts.

Why this matters now

Examination is the question regulators, auditors, and counterparties are starting to ask in this order:

Most compliance vendors answer the upstream sequence — identity, monitoring, interdiction. None of those answer the five questions above. Replay is the only primitive that does.

This page is Layer 5 of the stack. The signed verification report it emits is defined by §10 of the HATS Record v1 canonical specification, schema h33-signed-verify-report-v0.1. The replay bundle dropped in above is Layer 4.
HATS Record v1 spec →
Looking up a code from the verifier's JSON output? Every check, constraint, and reason code the verifier emits — HATS-V-CHECK-001 through HATS-G-008 and every other family — appears in the HATS code registry: 164 codes mirrored from the backend governance specification, with verifier-status annotations for each.

What this does NOT prove

Completeness

A bundle may be a truthful subset (e.g., date-range export). PASS does not assert "this is every event the case ever had."

Receipt validity

The 74-byte H33 receipts inside the bundle are validated by h33-verify, not the replay verifier. Use both for full coverage.

On-chain anchor

If the bundle was anchored to Polygon / Bitcoin / Solana, this playground does not look up the anchor transaction. Pair with the chain-specific verifier.

Verifier identity

A signed transcript proves some verifier with this public-key fingerprint produced this verdict. Whether that fingerprint belongs to whom you think it does is established out of band (allow-list, registry).

Run it yourself in the terminal

This playground is a thin wrapper. The same Rust core ships as a CLI — install once, verify anywhere:

$ curl -L -o h33-replay-verify https://h33.ai/dl/h33-replay-verify/v0.2.0/h33-replay-verify-v0.2.0-darwin-arm64
$ chmod +x h33-replay-verify
$ ./h33-replay-verify case-bundle.json --sign > signed-report.json
$ ./h33-replay-verify --verify-transcript signed-report.json
The verification layer should survive chain migrations, vendor changes, operator failure, and future infrastructure shifts. That is the standard the evidence and examination layer has to meet.
The bill is building a framework: identity → monitoring → interdiction → evidence → examination. Most firms are focused on the first three. The long-term winners may be the firms that provide the last two.