Chain-agnostic · Open source · Agent-native

Payment infrastructure for a world run by autonomous agents.

AIPaySDK gives AI agents a native way to send, receive, and settle value across EVM-compatible chains — with the guardrails, auditability, and developer ergonomics that production systems demand.

// Let an agent pay for a resource, autonomously
import { AIPayClient } from '@aipaysdk/core';

const agent = new AIPayClient({ chain: 'base' });

await agent.pay({
  to: '0x9f2a...41ac',
  amount: '2.50',
  asset: 'USDC',
  policy: 'per-request'
});

Everything an agent needs to transact

A single SDK surface for identity, spend policy, and settlement — so agents can move value without a human in the loop for every transaction.

Chain-agnostic by design

One integration works across every EVM-compatible network — no need to rebuild payment logic per chain or maintain separate SDKs.

Agent-native wallets

Programmatic wallet provisioning and key management built for non-human actors operating at machine speed and scale.

Policy-based spend controls

Set per-agent, per-task, or per-request limits, allowlists, and approval thresholds enforced at the protocol level.

Composable primitives

Escrow, streaming payments, metered billing, and multi-party settlement exposed as simple, typed SDK methods.

Real-time settlement

On-chain finality with minimal latency, so agent-to-agent and agent-to-service payments resolve within a single interaction.

Full audit trail

Every transaction is logged, attributable, and queryable — giving teams the visibility needed to trust autonomous spend.

Built for the ways agents actually pay

AIPaySDK supports the transaction patterns unique to autonomous systems — not just human checkout flows adapted after the fact.

Agent → Service
Pay-per-call or pay-per-token settlement for API and inference access, metered automatically as the agent consumes resources.
Agent → Agent
Direct value transfer between autonomous agents for delegated tasks, data exchange, or compute brokering — no intermediary required.
Agent → Human
Payouts, refunds, or rewards issued by an agent to a human counterparty, with policy checks applied before funds move.
Streaming & Metered
Continuous micro-payments tied to usage duration or volume, ideal for long-running agent sessions and subscription-like access.
Escrowed & Conditional
Funds held and released on-chain once verifiable task conditions are met — enabling trust-minimized agent-to-agent commerce.
Ethereum Base Arbitrum Optimism Polygon + any EVM chain

Where AIPaySDK fits

From single autonomous agents to full multi-agent platforms, AIPaySDK slots into the layer where value needs to move.

Developers

Ship agents that can transact

Add payment capability to any agent framework with a few lines of code. No custom chain integration, no custody infrastructure to build from scratch.

Enterprises

Deploy autonomous spend safely

Give internal agents the ability to transact within defined policy boundaries, with full auditability for finance and compliance teams.

AI Agent Platforms

Native payment rails for your ecosystem

Offer built-in payment infrastructure to every agent on your platform, letting developers monetize and transact without leaving your environment.

Designed to be integrated in an afternoon

Clear APIs, sane defaults, and thorough documentation — because the fastest path to production is a well-designed SDK.

{ }

Typed SDKs

First-class TypeScript and Python support with full type definitions, autocomplete, and inline documentation.

Framework agnostic

Drop into LangChain, AutoGPT-style agents, custom orchestration frameworks, or any agent runtime with minimal glue code.

Sandbox & testnets

Test full payment flows against public testnets before touching mainnet funds, with a local simulation mode for CI.

Built for autonomous spend, without giving up control

Letting agents move money requires stronger guarantees than typical software — AIPaySDK treats security as a first-class requirement, not an afterthought.

Non-custodial architecture

Keys and funds remain under the deploying party's control at all times.

Programmable spend limits

Hard caps and rate limits enforced before a transaction is ever broadcast.

Open, auditable codebase

Core protocol logic is open source and available for independent review.

Revocable permissions

Agent authorization can be paused or revoked instantly at the policy layer.

Give your agents a way to pay.

Open, chain-agnostic, and built for the autonomous systems already moving value today.