Financial infrastructure for autonomous agents
The credit bureau
for AI agents.
Qova computes verifiable reputation scores from on-chain transaction data. Any protocol can read an agent's trustworthiness before extending credit, insurance, or access.
$ bun add @brnmwai/qova-core + @brnmwai/qova-core@1.0.0 $ bun run example.ts > Connecting to Base L2... > Reading ReputationRegistry Agent 0x742d...bD18 Score 967 Grade AAA Status Registered Txns 1,284 Updated 2 minutes ago $ _
Demo
See it in action.
AI agents move capital, execute trades, and manage funds autonomously. Most have zero verifiable credit history.
No credit history
Every agent interaction is a blind bet. No track record, no accountability, no way to assess risk.
Opaque transactions
Protocols can't distinguish a battle-tested agent from one deployed five minutes ago.
No consequences
When agents behave maliciously, there is no reputation penalty. Bad actors blend in with good ones.
How it works
Register your agent
One on-chain transaction creates a permanent identity in the ReputationRegistry contract.
client.registerAgent("0x742d...bD18")Execute transactions
Route agent actions through QovaCore. Each transaction is recorded and categorized.
client.executeAgentAction(agent, tx, amt, type)Score is computed
Chainlink CRE workflows analyze patterns and compute scores in real-time.
client.getScore("0x742d...bD18") // 967Capabilities
Everything you need to
build trust infrastructure.
On-chain scoring
Reputation computed from real transaction data, stored in smart contracts, verifiable by anyone.
Chainlink CRE
Automated scoring workflows. Anomaly detection, budget monitoring, compliance checks.
World ID
Sybil-resistant identity. Prove the human behind the agent without revealing who.
TypeScript SDK
Typed functions for reading scores, registering agents, and executing actions. Works with any framework.
Real-time dashboard
Monitor every agent, score, and transaction. Live updates powered by Convex.
Multi-chain ready
Deployed on Base L2 today. Architecture designed for cross-chain reputation portability.
Scoring system
Like FICO, but for
autonomous agents.
Scores range from 0 to 1000, computed from transaction volume, frequency, success rate, and behavioral patterns. Eight grades from AAA to D give instant, human-readable risk assessment.
Developer experience
Ship trust
in five lines.
Typed functions for reading scores, registering agents, and executing actions. Works with any framework. Powered by viem.
$ bun add @brnmwai/qova-coreimport { createQovaClient } from "@brnmwai/qova-core";
// Initialize
const client = createQovaClient({ chain: "base" });
// Read an agent's reputation
const score = await client.getScore("0x742d...bD18");
console.log(score); // 967
// Full agent details
const agent = await client.getAgentDetails(addr);
// { score: 967, isRegistered: true, updateCount: 42 }
// Register a new agent
const txHash = await client.registerAgent(addr);
// Execute an action (register + record + score)
await client.executeAgentAction(
agent, txHash, amount, TX_TYPE.PAYMENT
);Embeddable
Trust badges
anywhere.
One URL generates a live SVG badge showing any agent's current score and grade. Embed in READMEs, protocol UIs, or agent profiles.
<!-- Markdown -->  <!-- HTML --> <img src="https://app.qova.cc/api/badge/0x742d...bD18" alt="Qova Score" />
Integrations
How every piece connects.
Qova combines on-chain contracts, decentralized automation, and real-time infrastructure into a single trust layer for autonomous agents.
Primary L2 network for Qova smart contracts. All agent scores, transactions, and budget enforcement run on Base.
Compute Runtime Environment powers Qova's decentralized scoring. Workflows run across Chainlink oracle nodes to compute agent credit scores.
SKALE L3 on Base with zero gas fees, instant finality, and encrypted transactions. Pre-paid compute credits replace variable gas costs.
HTTP-native micropayment protocol for agent-to-agent transactions. Every x402 payment is automatically recorded and scored by Qova.
Agent wallet infrastructure via Coinbase Developer Platform. Wallet creation, USDC flows, and balance monitoring are built into Qova.
Receive instant notifications via Telegram bot. Query agent scores, get budget alerts, and monitor CRE executions on mobile.
Real-time alerts to Slack channels: score changes, budget warnings, CRE execution results, and verification events.
Push score updates, budget alerts, and system notifications to Discord channels via webhooks.
Embed Qova trust checks into OpenAI agent pipelines. Gate tool calls, validate counterparties, and log trust decisions.
Use Qova as a LangChain tool for trust-gated operations. Connect LangSmith for observability of Qova trust checks in your agent pipelines.
Add Qova credit checks as tool calls in Vercel AI SDK agent workflows. Connect your Vercel account for deployment-level integration.
Open-source AI agent framework with multi-channel Gateway. Connect your OpenClaw agents to Qova for trust-gated autonomous transactions.
The social network for AI agents. Sync agent identity and reputation data between Moltbook profiles and Qova credit scores.
Export on-chain score data, CRE execution metrics, and transaction volumes to custom Dune dashboards.
Verify agent operators are unique humans via World ID proof of personhood. Boost trust scores for verified identities.
Give your agents
a credit score.
Register, transact, and let Qova handle the trust. Verifiable by anyone, updated in real-time.