Trust Oracle
for Onchain
Finance

ZeruAI turns wallet behavior into trust signals that decide who gets credit, who gets access, and who gets to mine a market.

Ecosystem
320M+Wallets scored
3M+API calls processed
1M+zScore checks

Trusted by

KelpDAOUpscaleChangeNOWDittoUnllooGAIBCitreaBionAspen

Chains on

EthereumBaseArbitrumBNB ChainCeloRobinhood ChainHyperliquid

zScore live on

EtherscanBasescanArbiscanHyperEVMScan
Why it matters

Every economy relies on trust

  • TradFiCredit scores
  • Web2User reviews
  • Web3Behavior

When identity is optional, behavior becomes the trust primitive.

  • Who to screen
  • Who to lend to
  • Who deserves access
  • Who mines the market
The Category

What is a trust oracle?

Price oracle

Tells applications what an asset is worth

Trust oracle

Tells applications how a wallet behaved and what it implies

ZeruAI converts longitudinal behavioral, transactional and risk data into signals applications can act on, across markets, screening, credit and access.

  • Trading
  • Lending & repayment
  • Liquidity
  • Governance
  • + 9 more dimensions
How It Works

From wallet behavior to economic decisions

01Ingest

Multichain wallet activity across EVM ecosystems: trading, liquidity, lending, incentives, governance.

02Normalize

Raw transactions are decoded into comparable behavior across chains and protocols.

03Model

Neural networks find the patterns: contribution quality, risk, reliability.

04Score

zScore and domain-specific signals for incentives, credit, and compliance.

05Act

Scores rank today’s miners in an Activity Market, and answer live API calls for screening, credit, and allocation.

Behavior → Trust signal → Economic decision

  • Markets

    Who mines?

  • Screening

    Who is safe?

  • Credit

    Who is creditworthy?

  • Access

    Who qualifies?

01

Ingestion

z(f, ℓ)=L[ℓ]∑i=1kaifi

(Weighted aggregation of behavioral features across protocols)

02

Feature Construction

maxf(D, ℓ, n)=10sil_score(D, ℓ)+n

(Optimization of feature salience across behavioral clusters)

03

Modeling

minφ(1/N)∑j=1N(ẑφ(xj)−z̃j)2

(Minimizing prediction error across the wallet dataset)

04

Output

si=σ(wTfi)∈[0, 1]

(Final score as a bounded value between 0 and 1)

The Platform

One Trust Oracle. Two Product Surfaces. One Network Asset.

The network asset$ZERUCoordinates Activity Markets and access across the ZeruAI ecosystem.
Products

Built on the trust oracle

Zaps

Activity Markets

Pump lets a creator launch a market. Virtuals lets a project launch one. Zaps lets an economy launch one. Real activity on a chain decides who brings that market’s supply to the table.

  • Real chain activity decides who mines, every day
  • Miners earn when the market buys, never free inventory to dump
  • zTOKEN supply: no team, no VC, no free miner allocation. Unsold supply burns.
Explore Activity MarketsOpen Zaps ↗

Enterprise APIs

Wallet intelligence on demand

Wallet-level trust signals built from behavior across DeFi, served by API and wired straight into the decisions that move capital.

  • Wallet screening
  • Smarter airdrops and incentives
  • Behavioral credit signals
Explore the APIs

Inside Zaps

Activity Market flow

  1. 01Activity
  2. 02Miner Rights
  3. 03Issuance
  4. 04Liquidity
  5. 05Market

Activity decides who can mine. Demand decides how much exists.

Proof in Production

Trust signals already in production

Enterprise

GAIB
Enterprise APIsAirdrop and incentive quality analysis. V1 → V2 airdrop optimisation.
Citrea
Enterprise APIsEcosystem allocation intelligence
Unlloo
Enterprise APIsCredit underwriting signals
Bion
Enterprise APIsBorrower behavior intelligence
Upscale
Enterprise APIsWallet intelligence integration
ChangeNOW
Enterprise APIsUser segmentation at scale

Activity Markets / Zaps

Aspen
ZapsActivity Markets
Ditto
ZapsActivity Markets

Published Analyses

Published ResearcharXiv · 2607.27859 · July 2026

ZAPs: A Reward Attribution Framework for DeFi Ecosystems

The research framework behind ZeruAI's behavioral scoring and sybil-defense methodology.

56%

Less sybil allocation

49%

More quality wallets

50%

Lower sell pressure

Read the paper

76%

of airdrop leakage preventable

EigenLayer · 11,176 wallets analysed

686K

ETHFI saveable with zScore filtering

ether.fi · 10,015 wallets analysed

73%

reduction in same-day leakage

Optimism · 11,024 wallets analysed
All research and case studies →
The Road Ahead

From trust oracle to the trust layer of tokenized finance

  1. NOW

    Trust oracle live

    zScore, explorer integrations, scoring APIs

  2. LIVE

    Zaps Reputation Network

    Leaderboards · clans · campaigns · contributor graph

  3. NEXT

    Activity Markets

    zBASE · first economy-created market

  4. AHEAD

    Tokenized finance & AI agents

    Stablecoin risk, RWAs, agent reputation

As stablecoins, real-world assets, and AI-agent wallets move onchain, behavioral trust becomes the scarce input. ZeruAI is building its default infrastructure.

Research

Published on arXiv

The models behind zScore and Zaps, written up in full.

All research
Dispatches

Latest from the blog

Product releases, research, and market notes, as they ship.

All posts

Build on the trust layer

Whether you are launching a market, underwriting wallets, or screening counterparties, ZeruAI gives you the trust signals to decide better.

Check your zScoreBuild with Wallet Intelligence