Abstract
Recent surges in LLM-driven intelligent systems largely overlook decades offoundational multi-agent systems (MAS) research, resulting in frameworks withcritical limitations such as centralization and inadequate trust andcommunication protocols. This paper introduces the Fetch.ai architecture, anindustrial-strength platform designed to bridge this gap by facilitating theintegration of classical MAS principles with modern AI capabilities. We presenta novel, multi-layered solution built on a decentralized foundation of on-chainblockchain services for verifiable identity, discovery, and transactions. Thisis complemented by a comprehensive development framework for creating secure,interoperable agents, a cloud-based platform for deployment, and an intelligentorchestration layer where an agent-native LLM translates high-level human goalsinto complex, multi-agent workflows. We demonstrate the deployed nature of thissystem through a decentralized logistics use case where autonomous agentsdynamically discover, negotiate, and transact with one another securely.Ultimately, the Fetch.ai stack provides a principled architecture for movingbeyond current agent implementations towards open, collaborative, andeconomically sustainable multi-agent ecosystems.