Hi HN, We built Blockrand, a randomness API for applications where “trust us” is not good enough (games, lotteries, simulations, on-chain/off-chain hybrids). Most “provably fair” RNGs fail under closer inspection: The server can influence outcomes The client can brute-force seeds Or verification only works after you already trust the operator Blockrand uses a double-blind model: The client commits to a seed hash (unknown to us) The server commits to entropy (unknown to the client) Final randomness is derived only after both are locked Anyone can independently verify the result later No party can bias the outcome, even if malicious. Why we built this We couldn’t find a randomness service that: Works off-chain Is cryptographically verifiable Doesn’t require trusting the operator Is simple enough to integrate into existing systems So we built one. What’s live today Simple REST API Deterministic, replay-verifiable results SDKs (JS, more coming) Open verification logic (no black boxes) What we’d love feedback on Threat model assumptions Attack vectors we may have missed Whether this solves a real problem for you API ergonomics (especially for game devs) This is early and opinionated. Happy to answer questions... https://blockrand.net GitHub: https://github.com/blockrand-api/blockrand-js