Internet of Agents: Fundamentals, Applications, and Challenges

TL;DR

The paper surveys the Internet of Agents (IoA), an agent-centric, scalable infrastructure designed to connect virtual and embodied agents for autonomous discovery, communication, and collaboration. It articulates a four-layer IoA architecture, key enabling mechanisms (capability discovery, task orchestration, semantic communication, consensus, and economics), and a spectrum of standardized protocols (MCP,A2A, ANP, AGNTCY, Agora) to enable inter-agent interoperability. It identifies open challenges in interoperation, security, privacy, governance, and ethics, and proposes research directions toward standardization, secure protocols, decentralized ecosystems, agent economies, and privacy-preserving interactions. The work aims to lay a foundational roadmap for realising large-scale, cross-domain, autonomous agent ecosystems with practical impact in smart cities, factories, healthcare, and beyond.

Abstract

With the rapid proliferation of large language models and vision-language models, AI agents have evolved from isolated, task-specific systems into autonomous, interactive entities capable of perceiving, reasoning, and acting without human intervention. As these agents proliferate across virtual and physical environments, from virtual assistants to embodied robots, the need for a unified, agent-centric infrastructure becomes paramount. In this survey, we introduce the Internet of Agents (IoA) as a foundational framework that enables seamless interconnection, dynamic discovery, and collaborative orchestration among heterogeneous agents at scale. We begin by presenting a general IoA architecture, highlighting its hierarchical organization, distinguishing features relative to the traditional Internet, and emerging applications. Next, we analyze the key operational enablers of IoA, including capability notification and discovery, adaptive communication protocols, dynamic task matching, consensus and conflict-resolution mechanisms, and incentive models. Finally, we identify open research directions toward building resilient and trustworthy IoA ecosystems.

Figures (9)

• Figure 1: Organization structure of this survey paper.
• Figure 2: Workflow of functional modules of virtual and embodied agents. Note that the blue text and blue dashed box are unique for embodied agents.
• Figure 3: An overview of IoA encompassing heterogeneous autonomous agents across diverse domains, with an illustrative example of cross-domain agent interactions. Consider two agents operating in distinct domains, e.g., domains 1 and 4. Initially, each agent registers with a proxy agent hosted on its respective domain gateway to facilitate agent discovery. Gateway proxy agents then broadcast the capabilities of all registered agents within their domain. Upon identifying a common task, the two agents initiate collaboration, facilitated by relay gateways that manage inter-domain communication. Throughout the interaction, gateway proxy agents coordinate and synchronize the agents' states, ensuring consistency and coherence across domains.
• Figure 4: General architecture of IoA, comprising four tiers from bottom to top: the infrastructure layer, agent management layer, agent coordination layer, and agentic application layer.
• Figure 5: An overview of IoA applications: (a) smart city, (b) smart home, and (c) smart factory.