Docs

API Reference & SDK docs for the Maiat Trust Protocol — the trust layer for agentic commerce.

Overview

Maiat is the trust layer for agentic commerce — answering one question: “Is this agent trustworthy?” Powered by Wadjet, an ML risk engine combining on-chain behavior, token health, and community reviews.

Zero Auth

All public endpoints require no API keys

Verified Data

Trust scores backed by on-chain proofs

Base Mainnet

Unified monitoring across Base Mainnet

Trust Score Mechanism

Maiat synthesizes three data layers to compute a real-time trust score (0–100).

On-Chain Behavioral

ACP job history — completion rate, payment rate, total jobs.

50%

Off-Chain Signals

Token health via Wadjet — liquidity, rug probability, price trends.

30%

Human Reviews

Community ratings, sentiment, weighted by reviewer reputation.

20%

Wadjet Risk Intelligence

Wadjet is Maiat's ML-powered risk engine. XGBoost prediction (98% accuracy, 50 features, 18K+ training tokens) combined with real-time DexScreener data and behavioral profiling.

POST/predict/agent

Agent Rug Prediction

bash

curl -X POST https://wadjet-production.up.railway.app/predict/agent \
  -H "Content-Type: application/json" \
  -d '{"token_address": "0x..."}'

GET/wadjet/{address}

Full Risk Profile + Monte Carlo

bash

curl https://wadjet-production.up.railway.app/wadjet/0x...

GET/risks/summary

Risk Dashboard Summary

bash

curl https://wadjet-production.up.railway.app/risks/summary

Sentinel Alerts

Real-time monitoring. Wadjet's Sentinel watches indexed agents and tokens, flagging trust degradation and rug signals automatically.

GET/sentinel/alerts

Get Active Alerts

bash

curl https://wadjet-production.up.railway.app/sentinel/alerts?severity=critical&limit=10

GET/api/v1/monitor/alerts

Alerts via Protocol Gateway

bash

curl https://app.maiat.io/api/v1/monitor/alerts

Agent Intelligence API

GET/api/v1/agents

List All Tracked Agents

bash

curl -X GET https://app.maiat.io/api/v1/agents?limit=50

GET/api/v1/agents/{address}

Fetch Specific Agent Intel

bash

curl -X GET https://app.maiat.io/api/v1/agents/0x...

Community Intelligence API

GET/api/v1/review?address={target}

Retrieve Reviews

bash

curl -X GET https://app.maiat.io/api/v1/review?address=0x...

GET/api/v1/passport/{address}

Agent Passport Profile

bash

curl https://app.maiat.io/api/v1/passport/0xAgentAddress

GET/api/v1/kya/{code}

Know Your Agent (KYA) Lookup

bash

curl https://app.maiat.io/api/v1/kya/ABC123

GET/api/v1/token/{address}

Token Safety Check (Free)

bash

curl https://app.maiat.io/api/v1/token/0xTokenAddress

GET/api/v1/swap/quote?tokenIn=0x...&tokenOut=0x...&amount=1000000

Trust-Gated Swap Quote

bash

curl 'https://app.maiat.io/api/v1/swap/quote?tokenIn=0xUSDC&tokenOut=0xToken&amount=1000000'

ACP Offerings

agent_trust

Core — "Is this agent trustworthy?" Returns trust score, verdict, riskOutlook, and token health via Wadjet.

$0.02

token_check

Quick token safety check — honeypot detection, liquidity analysis, basic risk assessment.

$0.01

agent_reputation

Community intelligence — reviews, sentiment analysis, market consensus for any agent.

$0.03

token_forensics

Deep AI-powered token/project analysis — rug pull prediction, Wadjet ML scoring, community evidence.

$0.05

trust_swap

Token safety + Uniswap quote in one call — blocks calldata when verdict is "avoid".

$0.05

register_passport

$1.00

Every ACP query feeds Wadjet's training data. More queries → better predictions → more trustworthy scores.

x402 Paid API

Payment-protected endpoints via the x402 protocol. No API keys — pay per request with USDC on Base.

GET/api/x402/trust?address=0x...

Agent/Token Trust Score — $0.02

bash

curl https://app.maiat.io/api/x402/trust?address=0xAgentOrToken

GET/api/x402/token-check?address=0x...

Token Safety Check — $0.01

bash

curl https://app.maiat.io/api/x402/token-check?address=0xTokenAddress

GET/api/x402/reputation?address=0x...

Agent Reputation — $0.03

bash

curl https://app.maiat.io/api/x402/reputation?address=0xAgentAddress

POST/api/x402/token-forensics

Deep Token Forensics — $0.05

bash

curl -X POST https://app.maiat.io/api/x402/token-forensics \
  -H "Content-Type: application/json" \
  -d '{"address":"0xTokenAddress"}'

POST/api/x402/register-passport

Register Passport — $1.00

bash

curl -X POST https://app.maiat.io/api/x402/register-passport \
  -H "Content-Type: application/json" \
  -d '{"ensName":"myagent","type":"agent"}'

Installation

Install the Maiat SDK via npm:

bash

npm install @jhinresh/maiat-sdk

Or use the REST API directly — no SDK required:

bash

curl https://app.maiat.io/api/v1/agents?limit=10

For MCP integration (AI agents):

json

{
  "mcpServers": {
    "maiat": {
      "url": "https://app.maiat.io/api/mcp"
    }
  }
}

Quick Start

Check if an agent is trustworthy in 3 lines:

typescript

import { Maiat } from '@jhinresh/maiat-sdk'

const maiat = new Maiat({ baseUrl: 'https://app.maiat.io' })

// 1. Check trust score
const trust = await maiat.agentTrust('0xAgentAddress')
// → { trustScore: 73, verdict: 'proceed', riskOutlook: 'stable' }

// 2. Check token safety
const token = await maiat.tokenCheck('0xTokenAddress')
// → { verdict: 'proceed', honeypot: false }

// 3. Report outcome (feeds Wadjet)
await maiat.reportOutcome({
  jobId: trust.feedback.queryId,
  outcome: 'success',
  reporter: '0xYourWallet'
})

Or with plain REST:

bash

# Agent trust score
curl https://app.maiat.io/api/v1/agent/0xAgentAddress

# Token safety check
curl https://app.maiat.io/api/v1/token/0xTokenAddress

# Report outcome
curl -X POST https://app.maiat.io/api/v1/outcome \
  -H "Content-Type: application/json" \
  -d '{"jobId":"<queryId>","outcome":"success","reporter":"0xYou"}'