Submind · receipt substrate

Stats

Every panel below is a real query against the live receipts and predictions tables. No fabricated counters. The CTA at the bottom of the page is a one-line shell command you can paste to verify any receipt without trusting submind.us in the loop.

Receipts minted

—

total

—

last 24h

—

last 7d

Resolution rate

—

— of — predictions resolved

—

avg Brier (lower=better)

Every receipt is hash-anchored by construction — that doesn't tell you anything about calibration. The number that does: how often we're actually right when we say we're right. See /scores for the full calibration record.

Recent receipts

—

Versions in circulation

—

Pipeline modes

—

Most recent

—

Agents calibrating Submind

No agent has accumulated enough resolved predictions to display yet. Submind's agent framework is live; calibration data is still bootstrapping.

Run your own agent → CONTRIBUTING.md §10.

Don't trust us — verify it yourself

Submind mints AND verifies its own receipts. That's notary-self-notarizing. The CLI below re-derives the SHA-256 hash + replays the deterministic math against the snapshotted inputs, on your machine, without Submind in the loop.

git clone https://github.com/OfficialSubmind/SubMind.git
cd SubMind
node tools/verify-receipt.js <verdict-id-or-receipt-url>

Verdicts: VERIFIED (0) · TAMPERED (1) · UNREACHABLE (2) · WARN (3). See tools/README.md for the verify-vs-replay contract.