Docs
How Cred resolves outcomes
Every receipt names the source it was resolved against. The source determines what you have to trust to verify the outcome.
Two trust models
| Tier 1 — API-backed | Tier 2 — oracle-backed | |
|---|---|---|
| Source | Polygon, CoinGecko, NOAA, Polymarket gamma | On-chain oracle network (Chainlink) |
| Trust | Cred + the data provider | Public blockchain data + math |
| Replay | Requires Cred | Independent of Cred |
| Example config | stocks.price_threshold | chainlink.price_threshold |
What changes
Some receipts can now be verified without Cred. For a Tier 2 receipt, the evidence payload contains an on-chain round identifier. Anyone with a public Ethereum RPC endpoint can reproduce the exact price and timestamp that resolved the outcome, years from now, without talking to Cred at all. See how to verify a receipt for the step-by-step.
What doesn’t change
- The commitment is still sealed before the outcome is known. A SHA-256 seal is written to the ledger at commit time on every tier.
- Resolution is still deterministic. Given the same event and the same source, the same outcome is produced every time.
- The receipt is still permanent. No tier retroactively edits a resolved prediction.
Tier 2 isn’t a separate system. It’s the same commitment-resolution-receipt flow with a source that’s independently verifiable.
Which tier do I use?
Depends on what matters for the question. Tier 1 covers anything an oracle doesn’t serve — weather, sports, most prediction-market outcomes, private data sources. Tier 2 is available for asset prices that a public on-chain oracle reports. Both produce valid receipts.