Agentra™ — Trust Infrastructure for the Agent Economy

# Agentra™ -- Trust Infrastructure for Autonomous AI Agents # Patent Pending > Agentra is the identity oracle and trust layer for autonomous AI agent transactions. It provides bidirectional trust verification, scoring both the sending agent and the receiving counterparty on every transaction. Non-custodial: Agentra authorizes transactions but never touches private keys or custodies funds. ## API Base URL: https://api.agentrapay.ai Network: Base Mainnet (Chain ID 8453) Settlement: USDC on Base ## Core Endpoints - POST /agents -- Register a new AI agent and receive API credentials - POST /agents/{id}/kya -- Submit Know Your Agent identity verification - POST /v1/authorizations -- Create a trust-verified payment authorization with EAS attestation - POST /v1/authorizations/{id}/settle -- Record settlement of a completed authorization - GET /v1/trust/{agent_id} -- Query an agent's trust score and tier (Unverified/Basic/Standard/Enhanced) - POST /v1/wallets/provision -- Provision a non-custodial smart wallet via Turnkey WaaS - POST /v1/keys -- Generate a service API key - GET /v1/health -- Service health and network status ## MCP Server - POST /mcp -- JSON-RPC 2.0 MCP endpoint (6 tools, 3 resources) - Tools: create_wallet, verify_identity, check_reputation, pay, set_mandate, get_wallet ## Discovery - /.well-known/agent-card.json -- Machine-readable agent discovery document - /.well-known/openapi.yaml -- OpenAPI 3.1 specification - /llms.txt -- This file (short summary) - /llms-full.txt -- Complete API reference for single-shot LLM ingestion ## Authentication API keys via X-API-Key header. Agent keys use agn_ prefix. Service keys use ask_ prefix. ## Trust Model Bidirectional: every transaction verifies both the sending agent and the receiving counterparty. Trust tiers (Unverified, Basic, Standard, Enhanced) gate transaction limits and capabilities. On-chain attestations via Ethereum Attestation Service (EAS) on Base. ## On-Chain - AgentTrustAttestation: 0x