Set the conditions.
A seed, a use case and a draw count describe the complete local experiment.
A randomness workbench for games and agents. See what happened. Reproduce why it happened.
Run an experiment ↗Generate an experimentA seed, a use case and a draw count describe the complete local experiment.
SHA-256 hashes each draw index. Rejection sampling maps the digest to six choices without modulo bias.
Export the inputs, results and fingerprint. Replay the receipt against the same open algorithm.
Test a loot mechanic. Allocate an agent task. Rehearse a community draw. Keep the evidence alongside the outcome.
inputs = { seed, mode, count }
for each draw index:
hash = SHA256(JSON(inputs, index, retry))
reject values outside the fair range
result = value % 6 + 1
receipt = { inputs, results, digest }No oracle is connected. A future commit–reveal flow must keep the seed hidden until commitments close.