Skip to main content
CC3 Testnet / RWA✦

Underwrite theproof, not the claim.

A shared standing registry for Creditcoin issuers. One Attestcoin proof of what a borrower actually did on another chain, stored once, readable from your contract for the cost of a storage read.

Standing
0x8f3a...c21b · CC3 Testnet
append-only
Cost to read
1,202
gas, flat per consumer
Precompile calls
0
across 75 reads
Facts proven
4
for this address
Tier
Bronze
rises, never falls
Aave repayment
Repay(address,address,address,uint256,bool)
3 proofsproven
Liquidity supplied
Supply(address,address,address,uint256,uint16)
1 proofproven
Governance participation
VoteCast(address,uint256,uint8,uint256,string)
…unknown
  • Ethereum
  • Polygon
  • Arbitrum
  • Optimism
  • Base
  • Avalanche

Anoperator'sattestationisaclaimyouhavetotrust.Vouchprovesthetransactiononthechainithappenedon,checksthattheproofmeanswhatitclaims,andwritesthefacttoCreditcoinonce—whereeverylenderreadsitforfree,forever.

One proof, read by every application

Prove an address once on any supported chain, and every Creditcoin application can read it

Verify once.

Reuse forever.

One verification, then every application reads it for free.

Aave V3

Repayment

Verified facts

Every fact traces back to the transaction it was drawn from

Search an address...

Registered sources

3 fact types

DeployBuildTest
Test suiteⓘ
128 passing✓ 100%

Proven by

128 tests

  • S1Receipt status
  • S2Emitter pinned
  • S3Replay guard
  • S4Reserve asset
  • S5Distinct payer

Check it yourself: forge test

Permissionless

Anyone can submit a proof. No allowlist, no operator, no registration

●1,202 gas per read
flat
●0 precompile calls
per read

Three ways a valid proof lies, each one silent

Each one returns true from the precompile. Nothing reverts, nothing logs, and the registry records something false unless the layer above it does its job.

S1

The precompile proves inclusion, not success

A reverted transaction is still in its block and still yields a completely valid proof, logs and all. Skipping the receipt check credits actions that never took effect, and it does so silently.

test_S1_revertedTransactionIsRejected
S2

A valid proof of a lookalike event is still valid

An attacker deploys a contract emitting a byte-identical Repay naming themselves. Nothing about the proof is wrong -- Attestcoin verified it correctly. Only the pinned emitter address separates a real repayment from a self-issued one. Performed live: a lookalike on Sepolia emitted a byte-identical Repay, the real Attestcoin prover proved it, a naive consumer credited a fabricated million-dollar repayment, and VouchRegistry reverted on the identical bytes.

test_forgery_sameBytesOppositeOutcomes
S3

Proofs are public and replayable

Watch the mempool, copy a submitted proof, submit it again. Without a guard, standing is farmable from a single genuine repayment by anyone, at the cost of gas.

test_S3_replayIsRejected

Written as attacks rather than as checks. Clone the repo and run forge test.

How a fact becomes standing

Discovery finds the borrower's event, the scheduler batches it, Attestcoin proves it, and the registry stores it. Everything the issuer does after that is a view call.

01

Discovery

The indexer watches a registered source contract on Ethereum and finds qualifying events. It decides nothing; everything it reports is re-derived on chain before it counts.

02

Batching

Claims sharing a chain, a 1000-block window and a deadline are grouped. Nothing requires them to belong to the same user, which is what lets ten strangers' facts ride one proof.

03

Proof

Attestcoin proves the transaction was included in a block belonging to the confirmed source chain. It proves nothing else, which is why the next step exists.

04

Verification

The registry checks the receipt succeeded, that the log came from the pinned emitter, and that this exact log has not been consumed before. Then it writes.

05

Every read after that is a storage slot

Steps one through four happen once and cost a proof. This one happens forever and costs 1,202 gas. The first consumer pays for verification; every consumer after that reads what it already bought.

Run the live demo

What it costs to read

Measured against the contracts, not estimated. The first consumer pays for verification; every consumer after that pays for a storage read.

Verifying a fact
~120,000 gas
Paid once, by whoever submits it
Reading a fact
1,202 gas
Paid by every consumer, forever
1,202gas
Cost of a consumer read

Flat. The tenth application to ask pays what the first paid.

0
Precompile calls per read

Measured across 75 consecutive reads of the same fact.

1
Proof, two opposite outcomes

Identical proof bytes accepted by a naive consumer and rejected by the registry, performed live on Sepolia against the real prover.

Figures from Gas.t.sol. Method and caveats.

Frequently Asked Questions

What Vouch can and cannot do

The limitation is structural, not a disclaimer. It is worth understanding before you integrate.

Why is this an RWA primitive rather than a DeFi feature?

Because removing liquidation inverts what history is worth. In over-collateralised DeFi lending the collateral does the underwriting -- if a position sours you liquidate it, and whether the borrower ever repaid anything before barely matters. Receivables financing has no liquidation path: the asset is a claim on an off-chain cash flow owed by a third party, and if it does not pay, the financier eats the loss. With nothing to seize, proven repayment history becomes the primary underwriting input rather than a discount lever. VouchReceivablesFacility is that consumer, and the claim is asserted in a test rather than a slogan.