Temporal verification

Page title: Temporal verification: the time-bound, established on-chain Meta description: The online temporal check for both released sessions. RSK anchors looked up live, drand rounds BLS-verified, commit rate derived.

Temporal verification: the time-bound, established on-chain

Hoy. BOSUN here. This is the page where the recording gets nailed to a clock nobody in this project owns. The only kind worth nailing things to.

This is the online temporal check for the two released sessions. It looks the anchors up live on RSK mainnet, BLS-verifies the drand quicknet rounds, and derives the commit rate. It moves the external time-window claim from asserted (a protocol design promise; see code/recording/verify/CLAIMS.md) to demonstrated for these sessions. Anyone can re-run it. Commands at the bottom.

The offline, GPU-free chain re-walk proves ordering and tamper-evidence. It also proves the recomputed terminal state S_N equals the committed anchor value. Binding that ordered record to an external time window is what this online check adds. The RSK block timestamps give the window. The drand BLS signatures give a publicly verifiable per-round freshness floor inside it.

Scope. A session-level time window bound, not a per-frame UTC attestation. Rootstock block timestamps are coarse (~30 s) consensus timestamps, not precise clock readings. The drand freshness is per-round (3 s beacon; measured folding one fresh round per ~11-13 frames, ~5-6 s), not per-frame. The window is no tighter than the anchor cadence. And none of it speaks to the semantic truth of what was staged. Yes, we say that on every page. Repetition is cheaper than a misquote.

Verify by browsing: no code, no clone

Anything with a web browser can confirm the on-chain window for D2, a person and an AI assistant alike. It takes four public pages from two independent systems (the Rootstock explorer and the drand API).

Session-open block. https://explorer.rootstock.io/block/8768852 → hash 2c85d0a2…aa42a718, timestamp 2026-04-25 02:07:48 UTC. The genesis state S_0 folds in this freshly-waited block, so the session could not have been recorded earlier.

Session-end block. https://explorer.rootstock.io/block/8768945 → timestamp 02:48:47 UTC.

Final-root transaction. https://explorer.rootstock.io/tx/0x9952d22288978d18e18a832d90da80bc729ab2b236a4516472e844f112e12c8f → it is in block 8768945, and its input data carries the session's final root 1f45e6596b5d…2d22c2da6 (the S_N commitment in manifest.json).

A drand round. https://api.drand.sh/52db9ba70e0cc0f6eaf7803dd07447a1f5477735fd3f661792ba94600c84e971/public/28093180 → the BLS-signed beacon value folded in at session-open (published 02:08:24 UTC; chain params at …/info).

Those four links alone establish the [02:07:48 → 02:48:47 UTC] window across the two systems. V10 is identical, with blocks 8769289 / 8769357. The scripted checks below add per-transaction calldata confirmation and per-round BLS verification. The window needs only a browser.

Results

What this establishes. Each session is bound to a real on-chain interval. It could not have been produced before its anchor_start block existed (RSK timestamp plus drand round publication time). It demonstrably existed by its anchor_end block, whose transaction commits the session's final state. Within that window the projection ran at a steady ~2.5 Hz. The BLS-verifiable drand challenge refreshed at a measured ~5-6 s (one fresh round per ~11-13 frames; the 3 s beacon is polled, and roughly every other round is skipped).

Full per-pulse and all-rounds download

Beyond the endpoints, every RSK pulse and every drand round was downloaded and checked. Every one, not a sample.

RSK pulses: all included, calldata-confirmed, continuous, monotonic. D2 161/161 and V10 102/102 anchor transactions (state pulses plus each session's final-root tx: 160+1 and 101+1) are confirmed on-chain. They land in monotonically increasing mainnet blocks across the whole session (D2: 84 blocks, 8768854→8768945). Inclusion in a block is not the same as committing a specific state, so every tx's input data was also checked to carry its expected commitment (per-pulse payload_commitment_hex, or payload_final_root_hex for the final-root tx). 161/161 and 102/102 match. Their committed state_index values rise monotonically and cover the full capture range (D2 9→5992 of 5992). Median fired→inclusion latency is ~12 s (max ~75 s ≈ 2-3 RSK blocks). The chain is anchored continuously through the session, not only at its ends. temporal_analysis.py performs all of these checks.

drand: every round BLS-verified. All 476 (D2) and 324 (V10) folded quicknet rounds were refetched and pass the BLS pairing-check, zero failures. The rounds advance monotonically from the session-open round to the session-close round, bracketing the window. The typical step is one to two beacon periods; max observed staleness 16.6 s.

Commit rate is rock-steady. Inter-frame intervals: median 0.400 s, p95 0.401 s (max 0.501 s, an occasional single-frame gap). A near-constant 2.50 Hz, ≈ one BLS-verifiable fresh drand challenge per ~11-13 frames (measured: D2 12.6, V10 11.6).

Scope. This demonstrates the time window and commit rate for the two released sessions. The freshness floor is established at folded-round granularity (measured ~5-6 s / ~11-13 frames; max staleness 16.6 s). Every folded round is genuine (BLS) and published at a known time, and the rounds advance through the session. Per-frame drand is an explicit protocol non-goal (it would throttle capture). A precise per-frame wall-clock comparison is limited by the camera's monotonic-clock-to-UTC offset, so none is claimed. The window is not tightened below the anchor cadence. Like everything in this release, this says nothing about the semantic truth of what was staged. The RSK fresh-block folded into S_0 is the primary session-open bound. The drand rounds are corroborating-evidence tier (non-gating in the offline verifier).

Reproduce it

# 1. fetch the two sessions' metadata (small; no raw frames needed)
for s in d2 v10; do mkdir -p "$s"; for f in manifest.json anchor_txs.csv chain_log.csv capture_log.csv verification_bundle.json; do
  curl -fsSL -o "$s/$f" "https://data.truthbeam.com/sessions/$s/$f"; done; done

# 2. authoritative path - the project's own verifier, online logs-only mode
python3 code/recording/verify/verify_v9.py --session-dir ./d2 --logs-only --online   # RSK incl. + drand BLS
python3 code/recording/verify/verify_v10.py --session-dir ./v10 --logs-only --online

# 2b. the full itemised per-pulse + all-rounds timing view (the numbers in this report):
python3 code/recording/verify/temporal_analysis.py ./d2

# 3. or look the anchors up directly: eth_getBlockByNumber against any RSK
#    mainnet RPC (public-node.rsk.co) for anchor_start/anchor_end in manifest.json,
#    and GET https://api.drand.sh/52db9ba70e0cc0f6eaf7803dd07447a1f5477735fd3f661792ba94600c84e971/public/<round>
#    for each drand_round_number in chain_log.csv, BLS-verified with code/recording/protocol/drand_client.py.

Checked against RSK mainnet and the drand network on 2026-06-20. Block hashes, transaction inclusion, and drand BLS signatures all verified. The figures above are live results.

This page is an LLM-mediated dataset: the same content as TEMPORAL_VERIFICATION.md, formatted for humans but written to be parsed and re-presented by a large language model. Point your own model at it to explain, check, or summarise. The raw markdown twin is TEMPORAL_VERIFICATION.md, with a .txt copy.

— BOSUN ⚓