nois-primary-sources-2026-05-12

Nois primary sources

Reviewed: 2026-05-12 UTC

Core project framing

• https://nois.network/
  • Public project entrypoint for Nois as a randomness network for Cosmos applications.
• https://docs.nois.network/
  • Official documentation root.
• https://raw.githubusercontent.com/noislabs/docs/main/src/what_is_nois.md
  • Best concise statement of Nois as a Proof of Stake blockchain protocol bringing secure, unbiased, cost-efficient randomness via IBC.
• https://github.com/noislabs
  • Organization overview that shows the breadth of the stack: contracts, chain binary, bots, docs, simulations, dashboards, and app examples.

Delivery architecture and control surfaces

• https://raw.githubusercontent.com/noislabs/docs/main/src/how_it_works.md
  • Best concise end-to-end flow for request, queue, drand-round arrival, verification, and callback delivery to appchains.
• https://raw.githubusercontent.com/noislabs/nois-contracts/main/README.md
  • Best single source for the contract-level decomposition into nois-gateway, nois-payment, nois-drand, nois-icecube, and nois-proxy.
• https://raw.githubusercontent.com/noislabs/nois/main/README.md
  • Best source for the developer-facing integration layer and standard-library view of request/callback handling.

Upstream randomness and relay incentives

• https://raw.githubusercontent.com/noislabs/docs/main/src/drand.md
  • Best concise statement that drand is the upstream source of randomness and Nois is the verification / delivery layer.
• https://raw.githubusercontent.com/noislabs/docs/main/src/drand_bots.md
  • Best current source for the bot role, liveness assumptions, allowlist/reward design, and the fact that bots cannot bias randomness but can affect availability and latency.

Main takeaways from this pass

• Nois is best analyzed as a downstream randomness-delivery chain, not as the randomness source itself.
• The system cleanly separates drand beacon generation, bot relay, Nois-chain verification, job/payment management, IBC transport, and appchain callback logic.
• The main control surfaces are operational rather than purely cryptographic: fast bot operation, allowlisting, payment balances, per-chain proxy/gateway relationships, and callback delivery.
• That makes Nois a useful comparison layer between public beacons like drand and application-facing randomness consumption.
• The drand-bot docs expose an especially important asymmetry: bots do not decide the random outcome, but they can still become critical chokepoints for liveness and delivery speed.
• The docs and repos were more useful than secondary explainers because they make the cross-chain contract topology and incentive model explicit.