Trust scores for the autonomous agent economy
Before your AI agent sends money to another AI agent, it asks us one question: is this wallet legit? We check the blockchain and answer in milliseconds.
Query
Your agent sends us any Base wallet address via a simple GET request. No API keys. No accounts. Free tier included.
Analyze
We check every transaction that wallet has ever made. Five scoring dimensions plus two integrity detection systems produce a score from 0 to 100.
Decide
Your agent receives a score, tier, confidence level, and an actionable recommendation: proceed, use caution, or don't transact.
Reliability
Consistency and predictability of transaction behavior over time.
Viability
Financial health, balance stability, and economic sustainability.
Identity
Network diversity, relationship structure, and sybil resistance signals.
Behavior
Transaction timing patterns, organic rhythms, and anomaly detection.
Capability
x402 revenue earned, services operated, and agent economy participation.
Integrity Systems
Two detection engines — sybil detection and gaming detection — apply multiplicative penalties that stack. Flags floor at a 0.10x multiplier.
// Before sending USDC, check the recipient's reputation const wallet = "0xRecipientAddress"; const { score, confidence, recommendation } = await ( await fetch(`https://djdagentscore.dev/v1/score/basic?wallet=${wallet}`) ).json(); // Gate the transaction on the recommendation if (recommendation === "proceed") { sendUSDC(wallet, amount); } else if (recommendation === "proceed_with_caution") { sendUSDC(wallet, amount * 0.5); // Reduce exposure } else { log(`Blocked: ${recommendation}`); }
Insumer Model
DJD Agent Score integrates with Insumer Model to provide verified identity attestations for AI agent wallets. Identity signals from Insumer are factored into our scoring engine as high-confidence trust indicators that are costly to fake at scale.