Skip to main content
AgentChain is the trust and settlement layer for autonomous actions on the EVM. Every agent action — API calls, token swaps, database mutations — produces a deterministic verification receipt anchored on-chain via EAS and pinned to IPFS. No blind trust required.

Quickstart

Submit your first verifiable task in under 5 minutes.

Architecture

How actions flow from policy → execution → verification → settlement.

Action Verification

How AgentChain verifies EVM state transitions deterministically.

API Reference

50 endpoints for operators, agents, and public consumers.

Core Concepts

Policies

On-chain rules that constrain what agents can do.

Approval Queue

Human-in-the-loop governance for high-value agent actions.

Attestations

Permanent EAS receipts for every verified action.

Disputes

Interactive bisection protocol for challenging invalid results.

For Operators

Submit tasks, manage API keys, track credits, and view verification receipts through the Operator API.

Submit a Task

Submit an action or inference task for verification.

Manage API Keys

Generate, list, and revoke operator API keys.

Guardrails & Approvals

Rate limits, spend caps, approval thresholds, and the kill switch.

Infrastructure

PRIOR Protocol

Cryptographic anchoring for task context and verification proofs on IPFS.

Governance

How AgentChainHelm manages upgradeable proxy contracts.