Supra’s core mechanism is a bundled oracle stack centered on DORA, where representative prices are computed from multi-source node inputs rather than simply forwarded from a fixed signer set.
The most useful analytical detail is the split between source collection, node-level medians, clan-level agreement, and final S-value publication. Supra’s docs and litepaper make that decomposition explicit enough to compare against other oracle architectures rather than treating it as a generic feed network.
The Tribe-Clan language matters because it signals that Supra is not only choosing reporters; it is shaping which subsets of nodes actually compute different commodities, which turns commodity assignment and anti-collusion policy into first-class design variables.
dVRF is doing double duty in the stack. It is marketed as a standalone randomness product for apps, but the data-feeds materials also use it to justify periodic task reassignment and anti-collusion properties inside the oracle network itself.
Supra exposes materially different downstream integration surfaces: push feeds for regular onchain updates, pull proofs for contracts that verify packaged oracle data on demand, and separate API / randomness service layers around the same broader stack.
Specific source signals worth retaining
oracles/data-feeds: best concise source for DORA, S-value, up-to-21-source aggregation, BFT framing, and the claim that dVRF helps preserve anti-collusion properties.
dora-distributed-oracle-agreement litepaper: best source for the core oracle problem framing, Tribe-Clan architecture, commodity sharding, node/source assignment via VRF, and why node-level medians are used before representative-value agreement.
pull-oracle: best source for the proof-carrying integration surface, especially the verifyOracleProof / verifyOracleProofV2 interfaces and the way proof generation can happen offchain before onchain verification.
push-oracle: best concise source for the contrasting distribution model where Supra publishes feeds onchain at high frequency for supported networks.
dvrf: best source for Supra’s randomness product as a separate subscription-and-callback control plane.
Why this entry belongs in the corpus
It strengthens the oracle cluster with a protocol that bundles consensus, anti-collusion assignment, multiple delivery modes, and standalone randomness rather than specializing in only one of those layers.
It is a useful comparison point for any oracle system that claims speed or decentralization while hiding who chooses sources, who packages proofs, who pays for updates, or how tasks are shuffled across operators.
The durable insight is that oracle power often sits upstream of the final signed number — in source assignment, committee topology, delivery mode, proof packaging, and scheduler/randomness design.