pattern

Credit Oracle Network

Propelling On-chain Economy

Credio streams machine learning model outputs by top data scientists and rating agencies directly into DeFi smart contracts in an automated, privacy-preserving and decentralized manner

animation placeholder

On-chain Issuers


Attract institutional capital

with credit ratings from top modellers

RWA Credit Investors


Analyse, execute and monitor RWA transactions,

real-time and automated

DeFi Protocols


Smart contracts can integrate with Credio

to automate their economic risk management

Logo
Logo
Logo
Logo
Logo
Logo
pioneers
A vision for decentralized zkML credit intelligence
WHAT

Credio supplies machine learning model outputs

directly to smart contracts

lead icon

Consumers: are DeFi issuers, investors or protocol DAOs that define risk parameter needs, set performance benchmarks, and establish rewards for Modellers.

lead icon

Modellers: build machine learning risk models and provide them as a service to earn revenues from ongoing use by Consumers.

lead icon

Validators: ensure models' integrity and quality by evaluating models against consumer needs, verifying zkML proofs and ensuring they meet the performance benchmarks.

lead icon

Oracle: the smart contracts that coordinate and incentivize the interactions between Consumers, Modellers and Validators to ensure the Credio provides high quality, decentralized model output feeds in a privacy-preserving manner.

Trusted by

Fasanara | Untangled

WHY

Programmable money

needs programmable creditworthiness.

Because money is credit.

lead icon

High quality inference feeds

  • - Built by top modellers and rating agencies
  • - Validated independently by third parties with 'skin in the game'
  • - Automated by smart contracts avoiding manipulation
lead icon

Decentralized, reliable and secured infrastructure

  • - Proven smart contracts end-to-end security
  • - Transparency through blockchain immutability
  • - Accessible and shared data infrastructure
lead icon

Privacy-preserving inferences through zkml technology

  • - Verifiable computation outputs without revealing the model workings
  • - Top modellers and rating agencies are incentivised to participate
  • - Cutting edge technology integrating AI/ML with blockchains
pattern2eco
approach
HOW

Credio incentivizes top data scientists to build models and generate inferences that are fed directly to on-chain applications

lead icon

Modeling challenge

  • - Consumers set out risk criteria/model outputs as challenges to Credio's decentralized community of modellers
  • - Modellers build and commit their models. They then compute unique test outputs against a challenge dataset along zkML proofs. Validators verify proofs and evaluate the inferences against performance benchmarks.
  • - Top performing model(s) are deployed and act as trusted modelling source, ready for consumption.
lead icon

Oracle operation

  • - A Consumer calls to request the model output feed through Credio as a custom feed from a particular model
  • - The request is then routed to a trusted model which computes and outputs risk parameters against the on-chain and off-chain data submitted by the Consumer.
  • - Inference is then provided to the Consumer along with zero knowledge proof. The entire process is seamless and programmable according to certain time-based or event-based triggers.
ROADMAP

Roadmap to a decentralized

risk intelligence ecosystem

lead icon

01: Data engineering and core model building

Engineer custom datasets combining both off and on-chain data and build machine learning models for a sample asset classes - working capital private credits

lead icon

02: Apply privacy-preserving technologies

Evaluate and incorporate zkML and OPML to both data inputs and model inferences

lead icon

03: Decentralizing modelling

Develop an ecosystem of modelers and output users (model consumers)

News and insights

DeFi yield landscape and the role of pegged assets

Comprehensive DeFi yield landscape including new restaking narrative and why peg stability of stable coins and staked ETH are essential for the growth and health of DeFi.

2024-10-07
5 Mins
Research

Use Credio to monitor stablecoin depegs in real time

We build machine learning models to predict depegs among 20 top stablecoins and use Credio to stream the results to blockchain in real-time

You can track the live results of Credio stablecoin depegging models <here>

2024-09-16
10 Mins
Presentation

Why RWA DeFi needs credit oracles?

Most RWA DeFi is ... credit: credit you extend to a centralised party (fiat-backed stablecoins like USDT and USDC or tokenised money market funds like BUIDL, ONDO or Benji) or you receive from a decentralised party (DAI) or a marketplace (like Untangled, Maple or Centrifuge)

2024-09-02
3 Mins
layer
layerlayer

Talk to an expert

Tell us about your credit intelligence needs or joining Credio's modelling community?