# AgentPay：A Programmable Payment Layer for Both Agents and Humans
Whitepaper Version: 0.7
Publication Date: January 2026
Author/Team: Agent Tech Research

## 0. Summary

**AgentPay is a stablecoin payment layer designed to enable agent-native payments across Web2, Web3, and AI-native environments.** Unlike traditional payment systems that treat payments as isolated transfers initiated by humans, AgentPay is built for autonomous agents, which supports programmable execution semantics, compliance-aware controls, and reliable settlement outcomes.

AgentPay provides a unified infrastructure for stablecoin payments and cross-chain coordination without compromising user self-custody. By productizing emerging agent-compatible payment standards such as **x402** through an SDK-first approach, AgentPay enables agents, businesses, and users to initiate and receive payments across fragmented blockchain ecosystems without embedding chain-specific logic into each integration.

In its initial phase, AgentPay focuses on enabling agent-native stablecoin payments through a modular payment stack, including agent-friendly payment interfaces, an execution decision hub, AgentPay facilitator, and cross-chain liquidity aggregator. Optional identity abstractions such as **Agent ID** can be used to improve attribution, policy control, and ecosystem-level accounting, while basic payment flows remain usable with standard endpoints such as wallets or email-based identifiers.

## 1. Motivation

### 1.1 The Rise of AI Agents as Execution Entities

AI systems are undergoing a structural transition. While early applications of AI primarily focused on decision support and analytics, recent advances have enabled agents to act as autonomous execution entities. These agents can initiate actions, interact with APIs, allocate resources, and complete tasks without continuous human intervention[^1].

As agents move from recommendation to execution, they increasingly require the ability to perform economic actions, such as paying for services, compensating other agents, or settling obligations across platforms[^2]. However, most existing payment systems are designed for human-initiated workflows, relying on manual approvals, static credentials, or account-based billing models. These systems do not scale naturally to autonomous, high-frequency, or programmatic execution[^3].

### 1.2 Stablecoins at Scale, Without Automation

In parallel, stablecoins have evolved into a widely adopted digital value layer for cross-border payments, on-chain settlement, and liquidity transfer across blockchain ecosystems. Their usage continues to expand across multiple contexts, serving as a common medium of exchange for digital economic activity[^4].

Despite this adoption, stablecoin payment flows remain largely human-centric. Users are expected to manage wallets, approve transactions manually, and navigate complex cross-chain processes. Liquidity remains fragmented across chains, and cross-chain transfers often involve high fees, operational risk, and multi-step execution. While stablecoins provide global reach and composability, they lack native support for automated and programmatic usage by autonomous systems[^5].

### 1.3 Structural Mismatch Between Agents and Stablecoins

The convergence of autonomous agents and stablecoin-based value transfer exposes a structural mismatch. Agents require embedded, programmable payment capabilities that can be invoked directly as part of execution logic, while stablecoins function primarily as passive value carriers without native mechanisms for agent-driven interaction.

Existing payment solutions typically address either automation or value transfer, but rarely both in an integrated manner. In addition, compliance, auditability, and risk management frameworks are not natively designed for agent-initiated transactions, creating further friction for adoption in production environments.[^6]

### 1.4 A New Infrastructure-Level Opportunity

This mismatch reflects a deeper shift in how economic activity is coordinated in digital systems. As agents become independent execution entities and stablecoins expand as a global settlement layer, payment systems built around manual, account-based workflows are no longer sufficient.

Rather than incrementally adapting legacy models, this shift creates the opportunity to design payment infrastructure explicitly for agent-driven, multi-chain execution, aligned with the emerging requirements of programmable and autonomous economic systems.

## 2. Project Vision

Our vision is to build an open infrastructure that enables **agent-native stablecoin payments** across multiple blockchains, supporting global and low-latency payment workflows while preserving user self-custody. The system is designed for an environment in which economic activity is increasingly executed by autonomous agents alongside humans, with funds remaining fully controlled in self-custodial wallets.

Core innovations include **a modular, agent-native payment architecture** and **the productization of emerging agent-compatible payment protocols**, such as x402. AgentPay decomposes payment execution into dedicated infrastructure components，including an **agent-friendly SDK**, an **execution decision hub**, on-chain **facilitator**, **liquidity aggregatator**, and optional identity abstractions called **Agent ID**, allowing payments to be initiated, authorized, routed, and settled as part of autonomous execution workflows. By turning protocol-level primitives into a production-ready payment layer, AgentPay enables programmable, agent-compatible payment execution across fragmented blockchain ecosystems, while supporting extensible settlement use cases such as cross-chain and multi-currency workflows.

The market opportunity is significant. Citi projects that total stablecoin market capitalization could reach between $1.9 trillion and $4.0 trillion by 2030, reinforcing the role of stablecoins as a global settlement layer[^7]. By addressing agent-native payments at the infrastructure level, AgentPay aims to reduce operational complexity, improve capital efficiency, and support scalable payment workflows while maintaining transparency, security, and user self-custody.

## 3. Market Overview

The stablecoin payment market is shaped by the convergence of multiple structural trends rather than a single technological driver. As illustrated in Figure below, a new class of payment demand is emerging at the intersection of three dimensions: stablecoins as a global value layer, cross-chain settlement across fragmented blockchain ecosystems, and the rise of AI-driven execution in economic workflows.
Stablecoins define the scale of value transfer, cross-chain settlement defines how that value moves across heterogeneous environments, and AI-driven execution introduces a demand-side shift in how frequently and programmatically payments are initiated. Together, these trends create infrastructure requirements that existing payment systems are not designed to address.

### 3.1 Structural Shifts in Stablecoin Payments

Stablecoins have become a core settlement medium for digital economic activity, supporting cross-border payments, on-chain settlement, and liquidity transfer across blockchain ecosystems. Assets such as USDC and USDT now underpin a growing share of global payment activity, with stablecoin transaction volumes reaching a record $33 trillion in 2025[^8].

At the same time, the nature of stablecoin usage is changing. Adoption is expanding beyond DeFi and speculative use cases into real payment flows, treasury operations, and institutional settlement. This shift reflects a transition from stablecoins as passive value holders toward stablecoins as active payment instruments.

Institutional participation is accelerating this trend. Major financial players are moving toward issuing regulated stablecoins, including euro-pegged initiatives by European banks and exploratory efforts around G7-linked stablecoins. Enabled by improved regulatory clarity in 2025, these developments indicate that future stablecoin supply will increasingly originate from traditional finance and fintech ecosystems[^9].

Besides, Visa Onchain Analytics further reveals a pronounced concentration within stablecoin usage. While total stablecoin transaction volume exceeded $58 trillion over the past 12 months, second-tier and long-tail stablecoins accounted for only approximately $74 billion, representing less than 0.2% of total on-chain activity[^10]. This imbalance indicates that the majority of stablecoin liquidity remains functionally underutilized outside of a small set of dominant assets.

### 3.2 Cross-Chain Fragmentation and Execution Constraints

Despite growing adoption, stablecoin payments remain fragmented across heterogeneous blockchain environments. Stablecoins circulate across multiple Layer 1 and Layer 2 networks, distributing liquidity across isolated execution domains and creating structural friction for settlement.

Cross-chain mechanisms are commonly used to bridge these environments, but they introduce significant risk and complexity. Between 2022 and 2025, bridge-related exploits resulted in losses exceeding $2.8 billion[^11]. In addition, cross-chain payment flows often involve multiple transactions, higher latency, and increased operational overhead.

This gap between the scale of stablecoin usage and the limitations of existing cross-chain infrastructure highlights the need for a unified execution layer capable of coordinating settlement across chains in a reliable and scalable manner.

### 3.3 AI-Driven Payments as a Demand-Side Amplifier

In parallel, AI-driven execution introduces a structural shift in payment demand. As autonomous agents increasingly initiate and coordinate economic actions, payments are becoming more frequent, programmatic, and embedded directly within execution logic.

Unlike traditional human-initiated payment flows, agent-driven payments are characterized by higher frequency and tighter coupling with automated workflows. Industry estimates suggest that AI-driven payments could cumulatively account for trillions of dollars in transaction volume over the coming decades. Rather than constituting a standalone market, this activity acts as a multiplier on existing stablecoin and cross-chain settlement infrastructure.[^12]

### 3.4 Optional Extension: On-Chain Forex Exchange

Beyond stablecoin-denominated payments, the emergence of multiple fiat-backed stablecoins introduces the potential for on-chain foreign exchange. While current stablecoin usage remains dominated by USD-denominated assets, increasing regulatory clarity and the issuance of non-USD stablecoins (e.g., EURC, JPYC) suggest the possibility of expanded multi-currency settlement on-chain.

Rather than constituting a standalone product category, on-chain foreign exchange is best viewed as an extension of stablecoin and cross-chain payment infrastructure, relying on the same underlying capabilities of liquidity coordination, execution routing, and compliance-aware settlement.

## 4. The AgentPay Stablecoin Payment Infrastructure

AgentPay is architected as a **programmable stablecoin payment layer** designed for agent-driven execution. Instead of treating payments as isolated value transfers, the system enables payments to function as programmable events that directly gate downstream actions and workflows. By providing a unified execution environment for stablecoin payments, AgentPay allows agents, applications, and services to initiate, receive, and settle payments in a way that is interoperable across networks while preserving user self-custody.

### 4.1 System Overview

This section describes how the AgentPay stablecoin payment infrastructure is organized at the system level. Rather than embedding payment logic across multiple applications or integrations, AgentPay structures payment execution as a unified layer composed of modular services with clearly defined responsibilities.

Each component within the payment layer participates in a distinct phase of payment initiation, decision-making, execution, or settlement, while remaining governed by consistent execution semantics and security assumptions. This modular structure allows the payment layer to support agent-driven workflows, heterogeneous execution environments, and evolving liquidity and settlement strategies without fragmenting developer interfaces or execution behavior.

#### 4.1.1 Payment Layer Architecture

The AgentPay payment layer is organized around five core components. Each component serves a clearly defined role and can evolve independently, while remaining governed by consistent execution and settlement semantics at the infrastructure level.
* **Application UI module**
The Application Interface provides the external interaction surface of AgentPay. It includes user-facing interfaces and developer-accessible endpoints through which payment intents are created, submitted, and monitored. This layer focuses on usability, request normalization, and integration convenience, without embedding execution logic or settlement assumptions.
Payment intents entering the system through this interface may originate from humans, agents, or hybrid workflows, and are passed downstream in a standardized format to ensure consistent processing across all execution environments.
* **AgentPay SDK**
The AgentPay SDK provides a programmatic interface for initiating and receiving payments within agent-native and application-driven workflows. Built with native support for emerging agent-compatible payment standards such as **x402**, the SDK allows developers and agent runtimes to express payment intents, receive execution outcomes, and consume settlement confirmations without directly handling blockchain-specific logic.
By encapsulating protocol interactions and execution semantics, the SDK enables autonomous agents to initiate payments as part of their execution logic, rather than treating payment as an external or manual operation.
* **AgentPay Execution Hub**
The Execution Hub functions as the **execution decision center** of the AgentPay payment layer. It evaluates incoming payment intents and determines whether, how, and under what conditions a payment may proceed. This includes applying programmable rules related to payment eligibility, limits, policy constraints, and execution authorization.
Importantly, the Execution Hub does not custody funds or perform settlement itself. Instead, it defines the execution outcome and governance logic that downstream components must follow. By centralizing decision-making at this layer, AgentPay ensures that payment behavior remains deterministic, auditable, and consistent across different integrations and environments.
* **AgentPay Facilitator**
The AgentPay Facilitator is a native settlement execution service responsible for carrying out approved payment instructions on-chain. Once a payment intent has been authorized by the Execution Hub, the Facilitator constructs, submits, and monitors the corresponding blockchain transactions across supported networks.
By operating as a dedicated settlement execution service, the Facilitator abstracts chain-specific transaction handling and confirmation logic from upstream components. This design allows AgentPay to support heterogeneous blockchain environments while maintaining a uniform notion of settlement finality at the payment layer.

* **Liquidity Aggregation Service**
The Liquidity Aggregation Service provides access to liquidity required for payment execution across chains and assets. It coordinates liquidity sources and defines how funds are sourced, routed, or temporarily advanced to ensure reliable settlement, particularly in cross-chain or low-liquidity scenarios.
Rather than embedding liquidity logic into execution or interface layers, AgentPay isolates liquidity coordination as a dedicated service. This separation improves capital efficiency, reduces execution risk, and allows liquidity strategies to evolve independently as the system scales.

Together, these components form a cohesive payment layer with the following properties:

- **Execution determinism:** Payment outcomes are governed by explicit decision logic rather than ad-hoc integrations.
- **Modularity:** Each service can evolve without fragmenting payment semantics or developer interfaces.
- **Self-custody preservation:** Users and agents retain control of funds; the payment layer coordinates execution without assuming asset ownership.
- **Agent-native design:** Payments are treated as first-class execution primitives within autonomous workflows.

This architecture allows AgentPay to function as a foundational payment layer for agent-driven economic activity, while remaining extensible to new protocols, networks, and execution models over time.

#### 4.1.2 Execution-Agnostic Interaction Coverage (ABC2ABC)

The AgentPay payment layer is designed to support execution across heterogeneous counterparties without embedding application-level roles or business assumptions. From the perspective of the payment layer, all participants are treated as execution endpoints defined by identifiers, authorization context, and settlement constraints.

This design enables native support for a broad spectrum of interaction patterns—including Agent-to-Agent (A2A), Agent-to-Business (A2B), Business-to-Consumer (B2C), Consumer-to-Agent (C2A), and other hybrid combinations—through a single execution and settlement framework. We refer to this capability as **ABC2ABC interaction coverage**.

Importantly, ABC2ABC is not implemented as a collection of specialized flows. Instead, it emerges from a role-agnostic execution model in which payment eligibility, settlement, and authorization are handled uniformly, regardless of the nature of the sender or recipient.

In practice, most existing payment systems are optimized for a limited subset of interaction types—such as consumer-to-business or business-to-consumer—and rely on off-layer logic or bespoke integrations to support agent-driven or cross-role payments. By contrast, AgentPay provides ABC2ABC coverage at the infrastructure layer, enabling agent-driven, business-driven, and user-driven payment flows to coexist within the same execution semantics.

#### 4.1.3 Composable and Evolvable Payment Layer

The AgentPay payment layer is designed as a composable infrastructure stack rather than a monolithic system. Each core component—SDK, Execution Hub, Facilitator, and Liquidity Aggregation Service—exposes well-defined interfaces and operates within clearly scoped responsibilities.

This design allows new execution policies, settlement strategies, liquidity mechanisms, or protocol integrations to be introduced without altering upstream application interfaces or breaking existing execution guarantees. Evolution occurs through the composition and configuration of components, rather than through ad-hoc extensions or bespoke integrations.

By structuring extensibility at the infrastructure level, AgentPay enables the payment layer to adapt to new agent frameworks, execution environments, and regulatory constraints over time, while preserving consistent execution semantics and developer experience.

### 4.2 Agent ID: Optional Identity Abstraction

AgentPay is designed to operate with flexible payment endpoints. Payments can be initiated and received using standard identifiers such as wallet addresses and email-based accounts, without requiring any dedicated identity primitive.

**Agent ID** is an optional, protocol-native identity abstraction that can be attached to a payment endpoint when stronger attribution, policy control, and ecosystem-level accounting are desired. It is designed to improve the ergonomics and governance of agent-native payments—without becoming a mandatory dependency for basic payment flows.

#### 4.2.1 Definition and Role

Agent ID is a persistent identifier that can represent an agent, application, service, or account context within the AgentPay payment layer. When enabled, it provides a stable reference for:
- attributing inbound payment activity to a consistent endpoint;
- applying configurable execution policies at the identity level;
- supporting usage-based accounting and incentive attribution;
- enabling future compliance configurations with minimal disruption to interfaces.

Importantly, Agent ID does not replace self-custodial authorization and does not alter the custody model of AgentPay. It functions as an identity overlay that enhances routing, policy enforcement, and attribution across heterogeneous environments.

#### 4.2.2 Flexible Payment Endpoint Addressing

Within AgentPay, payment endpoints may be expressed as:
- wallet addresses (EVM or non-EVM),
- email-based identifiers (account-linked endpoints),
- Agent IDs (when enabled).

These endpoint types are treated as interoperable inputs at the interface level. Resolution from an input identifier to a settlement target is handled by the AgentPay payment layer, allowing integrators to adopt Agent ID progressively without changing how payments are initiated.

#### 4.2.3 Issuance, Binding, and Optionality

Agent ID is designed to be optionally bound to an account or application context. A user may operate entirely without Agent IDs, using email-based endpoints or wallet addresses for both sending and receiving.

When an Agent ID is registered, it can be associated with one or more settlement destinations under defined rules (e.g., default wallet, rotating settlement address, or policy-scoped destinations). This approach allows identity continuity without forcing a single permanent address binding.

#### 4.2.4 Incentive Attribution and Usage Accounting

When incentive mechanisms such as Receive-to-Earn are enabled, attribution can be resolved at the Agent ID level to ensure incentives accrue to persistent endpoints rather than transient wallet addresses.

For deployments or flows that do not use Agent ID, the system can still account for payment activity based on available endpoint identifiers. Agent ID simply provides a stronger and more stable attribution anchor when needed—particularly for agent-operated services with recurring payment activity.

#### 4.2.5 Compliance Readiness and Policy Extensibility

Agent ID provides an additional abstraction layer where policy and compliance configurations can be attached in specific deployments. This may include identity-scoped acceptance policies, counterparty constraints, or risk controls, enabled under configurable settings of the payment layer.

### 4.3 Agent Execution Hub: Payment Decision and Authorization

The Agent Execution Hub is a core decision-making component within the AgentPay payment layer. It is responsible for evaluating payment intents, enforcing execution policies, and producing deterministic authorization outcomes that downstream components can rely on.

Rather than executing transfers or managing liquidity directly, the Agent Execution Hub defines *whether* and *under what conditions* a payment may proceed. It centralizes policy enforcement, compliance checks, and execution authorization, while delegating actual settlement and transaction submission to other components in the AgentPay stack.

By separating execution decisions from execution mechanics, the Agent Execution Hub enables AgentPay to operate as a scalable and extensible payment infrastructure for agent-driven workflows.

#### 4.3.1 Payment Intent Evaluation

The Agent Execution Hub evaluates incoming payment intents submitted through the AgentPay SDK or application interfaces. This evaluation ensures that each intent is well-formed, contextually valid, and suitable for further processing within the payment layer.

Intent evaluation establishes a consistent entry point for all payment flows, regardless of whether they originate from agents, applications, or services, and abstracts upstream systems from network- or asset-specific constraints.

#### 4.3.2 Policy and Limit Enforcement

The Agent Execution Hub enforces programmable policies and limits that define acceptable payment behavior. These policies may include rate limits, amount thresholds, execution scopes, and context-dependent constraints configured at the payment layer.

By enforcing policies at the decision layer rather than within settlement infrastructure, AgentPay ensures consistent execution semantics across heterogeneous networks and facilitator implementations.

#### 4.3.3 Compliance and Risk Screening

The Agent Execution Hub incorporates compliance and risk screening as part of the execution decision process. This includes pre-execution checks designed to detect abnormal, restricted, or high-risk payment activity before settlement is initiated.

Compliance logic is implemented as configurable decision rules within the hub, allowing different deployments to apply appropriate screening standards without modifying execution interfaces or custody assumptions.

#### 4.3.4 Execution Authorization

After a payment intent passes evaluation, policy enforcement, and compliance screening, the Agent Execution Hub determines whether execution is authorized. This authorization represents the canonical approval signal for downstream execution components.

Execution authorization is asset- and network-agnostic. It defines *permission to proceed* rather than *how execution is performed*, enabling settlement mechanisms and facilitators to evolve independently of decision logic.

#### 4.3.5 Authorization Outputs and Downstream Consumption

Upon authorization, the Agent Execution Hub produces a standardized execution authorization output. This output can be consumed by downstream components—such as facilitators or agent runtimes—to proceed with settlement or unlock subsequent actions.

In agent-driven workflows, this authorization acts as a deterministic execution gate. Downstream systems rely on verified authorization rather than interpreting raw transaction data, enabling clean separation between payment approval and task execution.

### 4.4 Receive-to-Earn and Incentive Flows

Receive-to-Earn is designed as a payment-level incentive trigger rather than a standalone economic program. It defines when incentives are activated and how rewards are attributed in response to valid payment activity within the AgentPay system.

The mechanism does not determine token issuance schedules, distribution ratios, or long-term value capture policies. These economic parameters are governed separately at the protocol level and are described in the Tokenomics section. This separation ensures that payment execution remains stable and predictable, while allowing incentive policies to evolve through governance.

#### 4.4.1 Receive-to-Earn: Incentivizing Payment Adoption

Receive-to-Earn is an incentive mechanism that rewards participants for receiving payments through AgentPay-native endpoints. When a payment is completed via supported interfaces such as x402-enabled workflows, a portion of protocol incentives is allocated to the recipient’s Agent ID.

Unlike transaction-based mining models that reward senders or arbitrage activity, Receive-to-Earn is structured to encourage acceptance and reuse of AgentPay as a payment rail. This approach is particularly relevant for merchants, service providers, and agent-operated services, where recurring inbound payment flows represent genuine economic activity rather than transient volume.

Rewards are calculated based on verifiable payment events and are settled at the Agent ID level, ensuring that incentives accrue to persistent identities rather than ephemeral wallet addresses.

#### 4.4.2 Agent ID–Based Attribution and Coefficients

All incentive distribution within AgentPay is resolved against Agent IDs. This allows the protocol to apply differentiated incentive coefficients based on identity characteristics without altering execution or settlement logic.

Certain Agent IDs may carry enhanced incentive parameters, including Super Agent IDs associated with merchants, branded identities, or scarce identifiers. These coefficients allow the protocol to prioritize strategic payment endpoints, improve merchant retention, and concentrate incentives on high-quality payment participants rather than raw transaction volume.

By decoupling incentive attribution from execution mechanics, AgentPay maintains flexibility to adjust incentive policies over time without impacting payment reliability or settlement performance.

#### 4.4.3 Relationship to Liquidity Incentives

Receive-to-Earn operates alongside liquidity incentives rather than replacing them. While liquidity providers are rewarded for supplying capital that enables payment execution, Receive-to-Earn targets the demand side of the network by encouraging payment acceptance and usage.

This dual-incentive structure supports both sides of the payment market:
Liquidity incentives ensure sufficient depth and execution reliability.
Receive-to-Earn incentives drive adoption, reuse, and endpoint stickiness.
Together, these mechanisms help stabilize network growth by aligning incentives across capital providers, merchants, users, and agent-driven services.
Incentive parameters and token issuance policies are governed by the Tokenomics framework described in Chapter 5.

### 4.5 Security, Compliance Readiness, and Risk Controls

This section describes the security architecture and compliance readiness of the AgentPay system. It outlines design principles, risk control mechanisms, and extensibility considerations, rather than jurisdiction-specific guarantees or regulatory representations. Security and compliance features are implemented to support evolving operational, governance, and regulatory requirements as the protocol matures.

#### 4.5.1 Execution Security and Key Management
All critical payment execution and settlement operations are protected through secure execution environments and policy-controlled signing mechanisms. Private key material associated with payment execution is isolated using trusted execution technologies or multi-party computation schemes, ensuring that no single system component can unilaterally access or misuse user funds. This approach supports single-signature user experiences while maintaining strong security guarantees across ABC2ABC payment flows.

#### 4.5.2 Operational Controls and Reconciliation

AgentPay enforces strict operational controls throughout the payment lifecycle. Every capital movement is cryptographically signed and recorded, producing immutable execution records and verifiable hash proofs. On-chain balances are continuously reconciled against internal accounting records, with automated alerts triggered in the event of inconsistencies or abnormal execution patterns. Additional safeguards such as transaction limits, time-based controls, and multi-level permission policies can be applied at the execution layer to further reduce operational risk.

#### 4.5.3 Auditability and Traceability

All payment execution, liquidity movements, and incentive-related events are recorded in a structured and queryable manner, enabling end-to-end traceability across chains and execution environments. These audit and traceability capabilities may be exposed through public or permissioned explorer interfaces (referred to as the **Agent Explorer**), allowing regulators, partners, or authorized participants to inspect cross-chain payment activity, execution records, and incentive attribution in a transparent and verifiable way.

#### 4.5.4 Compliance Readiness

While AgentPay is designed to operate in a self-custodial and permissionless manner by default, its architecture supports the attachment of compliance and risk attributes at appropriate abstraction levels, including Agent IDs and execution policies. This design allows the protocol to adapt to evolving regulatory requirements without disrupting core payment execution, liquidity coordination, or incentive mechanisms. Compliance controls can be progressively introduced as needed, enabling phased deployment across jurisdictions and use cases.

## 5. Tokenomics ($AGP)

### 5.1 The AGP Token Economy

The protocol issues a governance and utility token, AGP (AgentPay Token), which serves as the economic backbone of the AgentPay ecosystem. AGP is designed to align protocol usage, liquidity provision, and long-term value capture through a combination of governance rights, usage-based incentives, and a fee-driven buyback and burn mechanism.
AGP’s token model emphasizes participation in payment flows rather than speculative activity, ensuring that token incentives scale with real economic usage of the protocol.

**Core Functions:**
* **Governance**
AGP holders participate in protocol governance, including decisions on fee parameters, supported assets (stablecoins and fiat-backed tokens), protocol upgrades, and treasury management.
* **Liquidity Incentives (Liquidity Mining)**
A portion of the token supply is allocated to liquidity providers. By supplying stablecoins to AgentPay liquidity pools, LPs earn AGP rewards, supporting deep and reliable liquidity for payment, cross-chain settlement, and FX-related execution.
* **Fee-Driven Buyback and Burn**
A portion of protocol revenue generated from payment execution, cross-chain routing, and FX-related services is used to repurchase AGP from the open market and permanently burn it. This mechanism links long-term token value to sustained protocol usage rather than short-term emissions.
* **User Onboarding Rewards**
To encourage early adoption, new users interacting with the protocol—such as creating a self-custodial wallet, registering an Agent ID, or completing initial transactions—may receive limited AGP rewards as part of structured onboarding programs.
* **Receive-to-Earn (Usage-Based Incentives)**
AgentPay introduces a Receive-to-Earn mechanism that rewards productive participation in payment flows. Once a merchant, service provider, or agent registers an Agent ID, they receive programmable payment endpoints (e.g., accounts, APIs, or QR codes). When eligible payments are received through these endpoints, the receiver may earn AGP rewards.Receive-to-Earn is designed to incentivize real payment usage, not artificial volume. Reward distribution is governed by protocol-defined rules, including transaction validity checks, rate limits, and anti-abuse mechanisms. By rewarding receivers rather than only initiators, this mechanism improves user retention, encourages merchants to adopt AgentPay as their primary payment infrastructure, and embeds token incentives directly into recurring payment activity.

### 5.2 Fee Structure

AgentPay applies a two-tier fee model designed to balance system sustainability, execution efficiency, and abuse prevention, while keeping payment costs predictable for legitimate users.

* Processing Fee (Anti-Abuse Mechanism)
The processing fee is a minimal, protocol-level fee applied to certain payment execution requests to prevent spam, incentive exploitation, and automated reward farming.
This fee is not designed as a revenue source, but as an economic constraint that ensures only economically meaningful payment activity triggers execution and incentive logic. Processing fees may be waived, reduced, or dynamically adjusted for trusted flows, high-quality Agent IDs, or compliant usage patterns.

* Service Fee (Payment and Settlement Fee)
The service fee represents the standard cost of payment execution and settlement through the AgentPay infrastructure. It compensates liquidity providers and the protocol for routing, settlement coordination, liquidity usage, and operational overhead. Service fees are applied transparently and optimized through internal net settlement and routing efficiency.

* Fee Allocation
Service fees are distributed across protocol operations, and long-term sustainability mechanisms as governed by the Tokenomics framework. Processing fees primarily function as a system-level safeguard and are not intended to be a primary source of protocol revenue.

### 5.3 Sustainability

Long-term sustainability is supported through conservative reserve management and reinvestment strategies.

* **Stablecoin Reserve Management**
Portions of idle stablecoin reserves may be deployed into low-risk, audited DeFi protocols or real-world asset strategies (e.g., short-duration government securities), subject to governance oversight.
* **Yield Reinvestment**
Generated yield is directed to the protocol treasury to fund ongoing development, operations, insurance reserves, and counter-cyclical liquidity incentives, including enhanced Receive-to-Earn programs during periods of reduced activity.
* **Integrated Burn Mechanism**
Ongoing fee-based buyback and burn reinforces token scarcity, while usage-based incentives ensure that emissions remain tied to actual protocol activity.

## 6. Regulatory and Compliance Strategy

### 6.1 The Global Regulatory Landscape

We recognize that the future of stablecoins and digital payments depends on regulatory clarity. This protocol is designed with close attention to global regulatory trends, including the EU's MiCA (Markets in Crypto-Assets), guidance from the US SEC and CFTC, and recommendations from the Bank for International Settlements (BIS) on global stablecoin arrangements. Our goal is to build a framework that can operate compliantly across multiple jurisdictions.

### 6.2 Compliance-by-Design

The protocol adopts a "Compliance-by-Design" philosophy, embedding compliance tools into the infrastructure rather than adding them as an afterthought.
* **Compliance Configuration Layer:** AgentPay is designed to support configurable counterparty policies in specific deployments, including optional identity-based access controls and jurisdictional requirements. These controls can be enabled where needed without changing core payment interfaces or custody assumptions.
* **Anti-Money Laundering (AML) Tracking:** Leveraging the transparency of the blockchain, we will partner with leading on-chain analytics firms (e.g., Chainalysis, TRM Labs) to monitor suspicious flows and provide tools to comply with AML and CFT requirements.
* **Purpose Bound Money (PBM) Protocol:** For institutional and B2B use cases, the protocol will explore integrating PBM concepts, allowing senders to specify the use of funds (e.g., "salary payments only"), which greatly enhances transparency and control for institutional adoption.
* **Stability Assurance:** The protocol will enforce strict asset admission criteria, prioritizing stablecoins with transparent, verifiable reserve reports and strong market reputations to mitigate de-pegging risk.
* **Audit Reports & Proof-of-Reserves (PoR):** All smart contracts will undergo rigorous audits by top-tier security firms. Furthermore, for any centralized custody portions, regular Proof-of-Reserves (PoR) audits will be provided to ensure all assets are fully backed.

## 7. Risk Assessment and Mitigation

Agent Pay operates at the intersection of multi-chain infrastructure, autonomous agents, and programmable settlement. While the system is designed with security, modularity, and scalability in mind, it inevitably faces technical, execution, and market-related risks. This section outlines the primary risk categories relevant to Agent Pay and the corresponding mitigation strategies adopted at the system and operational levels.

### 7.1 Risk Overview and Classification
#### 7.1.1 Technical and Execution Risks

* **Smart Contract Vulnerabilities**
Despite rigorous development practices and multiple rounds of security audits, any complex smart contract system carries inherent risks of unforeseen vulnerabilities. Logic flaws, edge cases, or unintended interactions with external contracts may result in unexpected behavior or financial loss.
* **Chain Congestion and Layer-1 Risk**
Agent Pay relies on multiple Layer-1 and Layer-2 blockchains for payment execution and settlement. Extreme network congestion, prolonged downtime, consensus failures, or hard forks on a source or destination chain could lead to transaction delays, increased execution costs, or settlement failures.
* **Agent Execution and Authorization Risk**
In agent-driven payment systems, execution risk extends beyond traditional smart contract behavior. Autonomous agents may initiate payments at high frequency, across multiple services, or under compromised control conditions. Without clearly defined execution boundaries and authorization constraints, such activity could result in unintended fund movements, excessive spending, or abuse of delegated payment permissions.

#### 7.1.2 Market and Adoption Risks

* **Stablecoin De-pegging and Asset Volatility**
The effectiveness of Agent Pay depends in part on the price stability of supported stablecoins. Significant de-pegging events could reduce the effective value of settled assets, disrupt payment flows, and trigger liquidity stress or withdrawal behavior.
* **Liquidity Risk**
Under extreme market conditions, simultaneous withdrawals by liquidity providers may lead to increased slippage, reduced execution efficiency, or temporary interruptions in specific settlement routes or payment paths.
* **Adoption and Competitive Pressure**
The market already contains numerous payment, settlement, and cross-chain solutions. Agent Pay must demonstrate sustained advantages in security, cost efficiency, composability, and agent-native execution to overcome user, developer, and merchant inertia and achieve broad adoption.

### 7.2 Mitigation Strategies

To address the risks identified above, Agent Pay adopts a multi-layered mitigation strategy that combines protocol-level safeguards, execution controls, and operational resilience mechanisms.
* **Audits and Bug Bounty Programs**
Multiple rounds of independent external security audits are conducted prior to mainnet deployment. These are complemented by an ongoing bug bounty program designed to encourage continuous review and responsible disclosure from the broader security community.
* **Multi-Signature Controls and Trusted Execution Environments**
Critical protocol operations and treasury assets are protected using multi-signature wallet schemes and Trusted Execution Environments (TEE), reducing single points of failure and minimizing private key exposure.
* **Insurance and Reserve Mechanisms**
A portion of protocol revenue may be allocated to insurance or reserve funds intended to mitigate user losses arising from unforeseen technical failures or extreme operational events.
* **Dynamic Risk Parameters and Circuit Breakers**
Agent Pay continuously monitors on-chain conditions, liquidity health, and stablecoin pegs. When elevated risk conditions are detected, the system can dynamically adjust operational parameters, pause specific routes, or trigger circuit breakers to limit further exposure.
* **KYA-Ready Execution Controls**
To mitigate execution risks introduced by autonomous agents, Agent Pay enforces programmable payment constraints and authorization scopes at the payment and settlement layer. These include transaction limits, frequency controls, and auditable execution records linked to verifiable agent identities. Rather than interpreting agent intent or behavior, Agent Pay provides deterministic execution boundaries, allowing upstream agent platforms and compliance systems to retain full control over KYA logic.
* **Stress Testing and Simulation**
Extensive stress testing is conducted in testnet environments to simulate extreme market volatility, liquidity shocks, and chain congestion scenarios, ensuring system robustness prior to production deployment.

## 8. Vision and Roadmap

### 8.1 Development Phases

* **Phase 1 (2026 Q1 - Alpha Launch)**
* Release x402 integration payment.
* Release initial SDK for developers.
* Begin security audits and stress testing.
* Release Agent ID feature.
* **Phase 2 (2026 Q3 - Mainnet Launch)**
* Official mainnet launch, supporting major EVM and Non-EVM chains.
* AGP Token Generation Event (TGE) and launch of the liquidity mining program.
* Airdrop Events.
* Release full API/SDK for Agents and UCP.
* **Phase 3 (2027 - Global Expansion & AI Integration)**
* launch agentpay-facilitator allow multichain and multi-stablecoin autonomous payments for AI Agents.
* Launch institutional-grade compliance modules.
* Expand support for non-fiat stablecoins (e.g., commodity tokens).
* Establish global partnerships to promote the protocol as a standard for ABC2ABC cross-border payment.

### 8.2 Ecosystem Expansion

Our ultimate goal extends beyond simple stablecoin payment.
* **The AI-Driven Economy:** Through x402 and AgentPay, we will power the forthcoming AI Agent economy, enabling AI to autonomously manage finances, pay for compute resources, and participate in DeFi.
* **The Bridge for Web2 and Web3:** Our SDK and API will enable traditional FinTech and Web2 enterprises to seamlessly access on-chain finance, enjoying the efficiency and low cost of stablecoin payments without handling the underlying blockchain complexity.

### 8.3 Conclusion

The Multi-Chain Stablecoin Payment Protocol is not just a fix for the current fragmented payment system; it is the blueprint for the next generation of financial infrastructure. By connecting siloed blockchains and granting native payment capabilities to AI, we are committed to enabling a new era of global finance that is truly instant, inclusive, and efficient.

## 9. Appendix

### 9.1 Glossary

* **Cross-Chain:** The technology that enables the transfer of assets or information between two or more independent blockchains.
* **Stablecoin Peg:** The mechanism by which a stablecoin (e.g., USDC) maintains its value relative to a target asset (usually 1:1 with the US Dollar).
* **AgentPay:** A stablecoin payment infrastructure centered on a Programmable Payment Layer that governs payment eligibility, on-chain settlement outcomes, and execution authorization for agent-native workflows.
* **x402 Protocol:** An emerging standard designed to provide AI Agents with native, on-chain payment and settlement capabilities.
* **Agent ID:** An optional, protocol-native identity abstraction that can be associated with payment endpoints in AgentPay to support attribution, policy enforcement, and incentive accounting.
* **On-Chain Forex:** A system for the direct, real-time exchange of different national fiat stablecoins (e.g., USDC for EURC) on the blockchain.
* **ABC2ABC:** The overarching architecture supporting Agent-to-Agent (A2A), Business-to-Business (B2B), Consumer-to-Consumer (C2C), and all hybrid
combinations.

### 9.2 References

[^1]: arXiv. *AI Agents: Evolution, Architecture, and Real-World Impact* (2025).
https://arxiv.org/html/2503.12687v1

[^2]: Wikipedia. *Agentic commerce* (accessed 2025).
https://en.wikipedia.org/wiki/Agentic_commerce

[^3]: arXiv. *Unlocking AI Agents’ Potential Through Market Forces* (2025).
https://arxiv.org/html/2501.10388v2

[^4]: Bank for International Settlements (BIS). *Advancing in Tandem: Results of the 2024 BIS Survey on Cryptoassets* (2024).
https://www.bis.org/publ/bppdf/bispap159.pdf

[^5]: Fireblocks. *Global Insights: Stablecoin Payments & Infrastructure Trends* (2025).
https://www.fireblocks.com/report/state-of-stablecoins

[^6]: Google Cloud. *Announcing the Agents-to-Payments (AP2) Protocol* (2025).
https://cloud.google.com/blog/products/ai-machine-learning/announcing-agents-to-payments-ap2-protocol

[^7]: Citigroup Global Perspectives & Solutions. *Stablecoins 2030* (2025).
https://www.citigroup.com/rcs/citigpa/storage/public/GPS_Report_Stablecoins_2030.pdf

[^8]: Bloomberg. *Stablecoin Transactions Rose to Record $33 Trillion, Led by USDC* (2026).
https://www.bloomberg.com/news/articles/2026-01-08/stablecoin-transactions-rose-to-record-33-trillion-led-by-usdc

[^9]: Reuters. *European Banks Led by BNP, ING Push Ahead on Euro Stablecoin Plan* (2025).
https://www.reuters.com/business/finance/group-european-banks-announce-euro-stablecoin-plans-2025-12-02/

[^10]: Visa. *Visa Onchain Analytics: Stablecoin Transactions Dashboard* (accessed 2025).
https://visaonchainanalytics.com/transactions

[^11]: Chainlink. *Cross-Chain Security Report* (2025).
https://chain.link/education-hub/blockchain-interoperability

[^12]: McKinsey & Company. *The Agentic Commerce Opportunity: How AI Agents Are Ushering in a New Era for Consumers and Merchants* (2025).
https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-agentic-commerce-opportunity-how-ai-agents-are-ushering-in-a-new-era-for-consumers-and-merchants