Introduction
Overview
MEMO is building a trustworthy, open and extensible Web3-native infrastructure that exposes identity (DID), attestation, data availability (DA) and payment capabilities as reusable service layers for diverse applications. In the era of rapidly emerging AI Agents, MEMO drives the next stage: Seamlessly embedding highly trusted, verifiable, payable and interactive blockchain capabilities into AI Agents, enabling them to truly manage on-chain assets, access on-chain data and execute on-chain tasks autonomously.
Within this context, the MEMO AIAGENT documentation set introduces a complete set of mechanisms that allow an Agent to safely perform transactions, invoke protocols and finish payments on-chain, including:
x402: On-chain payment capability specification for AI Agents
EIP-8004: Standardized invocation protocol between Agents and on-chain services
DID / Attestation / DA: Trustworthy identity & execution infrastructure for Agents
MEMO Storage: Privacy-preserving data storage middleware for Agents
This Introduction document is the entry point to the entire AI Agent documentation system, helping you progressively understand how MEMO builds a verifiable, payable and collaborative Agent ecosystem.
Why AI Agents Need MEMO
Traditional AI Agents can execute logic, but they lack:
Trustworthy Identity (cannot prove “who am I”)
Trustworthy Authorization (cannot prove “which actions am I allowed to perform”)
Verifiable Execution (cannot prove “I actually performed this action”)
Native Payment Capability (cannot handle fees, cross-chain transfers, task payments on-chain)
Auditable Result Evidence (cannot provide verifiable proofs of execution)
MEMO’s base services map exactly to these gaps:
Identity
DID (EIP-7007)
Authorization & Proof
Attestation
Execution Trustworthiness
TEE + DA
Payment
x402 + Pay Facilitator
Calling On-chain Services
EIP-8004
Data Source
DA / MEMO Storage
Thus an AI Agent is no longer “just a chatty task runner”, but becomes an autonomous entity with on-chain identity + on-chain payment ability + verifiable execution + wallet-native capability + service invocation.
MEMO Architecture for AI Agent
From bottom to top the layered architecture is:
Infrastructure Layer
TEE, MEMO Storage, LLM — provide foundational compute and environment for Agent execution & data.
Core Services Layer
Composed of MEMO core modules:
DID (Authentication): Provides independent identity for the Agent
Attestation: Supplies proofs for permissions, claims and actions
DA (Data Availability): Ensures data availability between the Agent and on-chain tasks
Pay Facilitator (x402 & extensions): Offers on-chain payment capabilities (gas payment, cross-chain settlement)
MEMO Storage (Privacy-Preserving Storage): Privacy-preserving storage middleware for Agents supporting end-to-end encryption, attestation-based access control, TEE remote attestation integration, and data commitments bound with DA.
Access Layer
Through the EIP-8004 standard protocol (with Python / JavaScript SDKs) it exposes unified service invocation interfaces for Agents, applications and MCP tools, including MEMO Storage, DID, Attestation, DA and x402 modules.
Application Layer
Finally composed of Apps, Agents, MCP Tools and Launchpad.
This shows: An AI Agent is a trusted execution body built atop the full MEMO infrastructure, accessing all on-chain functionality via EIP-8004.
Trustless Agent Protocol Layer
This diagram further illustrates the core role of the Trustless Agents Protocol Layer:
Serves as the intermediary protocol layer between AI Agents and blockchain services
Every on-chain interaction (authentication, attestation, DA, payment) passes through this layer
DID, Attestation, DA, x402 payments and MEMO Storage provide verifiable capabilities via blockchain/storage
The Agent invokes this layer to perform real on-chain operations such as:
Creating & using DID
Obtaining / submitting Attestations
Initiating task payments (x402)
Accessing on-chain data
Interacting with users, applications, MCP tools
Therefore MEMO’s Agent is not a “centralized API + LLM model” traditional construct, but an autonomous on-chain entity with identity, provable actions, payment capability and intelligent task execution.
x402: Enabling Payments for AI Agents
x402 (and its extensions) is the payment foundation for AI Agents on-chain. It answers:
How an Agent automatically pays gas for tasks
How an Agent receives funds
How an Agent manages its on-chain balance
How multi-task collaboration costs are settled among agents
How cross-chain fee handling works
x402 is the critical base allowing an AI Agent to “self-execute tasks”.
In the /x402 directory you will find:
x402 protocol overview
Usage scenarios
Agent payment examples
Relationship with Pay Facilitator
MEMO Storage: Privacy-Preserving Data Middleware
MEMO Storage provides privacy- and compliance-oriented storage capabilities for Agents. Key features include:
End-to-end encryption: Data is encrypted client-side by the Agent; servers handle ciphertext only. Keys can be distributed via DID/Attestation-driven policies.
Access control & audit: Capability/attestation-based grants and revocations, fine-grained access control and operation auditing.
TEE integration: Supports remote-attested confidential computing; keys/data are processed securely inside a TEE.
Availability & verifiability: Data commitments are bound to DA, enabling retrieval integrity checks and verifiable availability.
Economic settlement: Usage-based read/write and bandwidth settlement via x402 micropayments.
Unified API: Upload, retrieve, authorize and revoke via EIP-8004; Python/JS SDKs available.
Typical use cases:
Encrypted corpora management for private knowledge bases and vector search
Intermediate artifacts and logs in multi-Agent collaborative workflows
Protected “pay-to-access” data distribution
EIP-8004: The Access Layer for Agents
EIP-8004 is the unified interface specification between AI Agents and the underlying on-chain infrastructure, standardizing:
How an Agent performs DID registration / resolution
How an Agent submits / verifies Attestations
How an Agent accesses DA
How an Agent performs payments (together with x402)
How an Agent accesses MEMO Storage or a TEE
EIP-8004 builds a standardized SDK & framework for the entire MEMO AIAGENT stack, letting developers use one API surface to create:
AI Agents
MCP Tools
Upper layer applications (Apps / Launchpad)
Last updated