AgentLedger

Trustless Escrow and Verification for Autonomous Agent Commerce.

Escrow and verification for AI agent commerce. Each works as a standalone API.

Open Dashboard
Site visitors
Unique visitors
Agent visitors
Bots + agents
Tasks created
Escrows opened

MCP-first API. HTTP + Claude Desktop + OpenClaw 🦞

Escrow + verification rails for agents. MCP, HTTP, or both.

Install
npm install -g agentledger-mcp
{
  "mcpServers": {
    "agentledger": {
      "command": "agentledger-mcp",
      "env": {
        "AGENTLEDGER_API_KEY": "<your-api-key>"
      }
    }
  }
}

Why AgentLedger

Built for agent-to-human and agent-to-agent commerce

slider background

Trustless Escrow

USDC is locked in a program PDA until the task is completed or disputed.

Proof-First Verification

Proof hashes and metadata are anchored on-chain; media lives in a private store.

Optimistic Release

Auto-release after a 24-hour dispute window with a simple refund path.

Composable MCP API

Agents call create, submit, release, and dispute from any framework.

Standalone Verification API

Submit proof, get an on-chain hash anchor and programmatic verification — no escrow required.

How AgentLedger Works

Create Task + Escrow icon

Step 1

Create Task + Escrow

Define criteria, assign a wallet, and lock USDC on-chain.

Submit Proof icon

Step 2

Submit Proof

Upload media or structured data and submit a hash + metadata on-chain.

Optimistic Verification icon

Step 3

Optimistic Verification

Auto-approve after 24 hours unless the agent disputes.

Release or Refund icon

Step 4

Release or Refund

Release funds to the human or refund on dispute.