SpoonOS announced as an operating system for AI agents on blockchain
Neo has announced SpoonOS, an agentic operating system design to support AI agents in decentralized environments. The platform provides the infrastructure for developers to build, coordinate, and scale on-chain LLM-powered applications. The project is aimed at serving as the foundation for a broader “sentient economy,” where autonomous agents interact, learn, and generate value across Web3 ecosystems. Though blockchain agnostic by design, Neo will be supported as a first-class citizen. Platform development SpoonOS is being incubated by Neo Global Development with the strategic goal of expanding the range of use cases within the Neo ecosystem using on-chain agents. Neo founder Da Hongfei said: With SpoonOS, we’re taking a bold step forward. Backed fully by Neo’s funding, resources, and technology, our mission is to build a universal agentic operating system for web3, and in doing so elevate Neo’s long-standing vision of a Smart Economy into a truly Sentient Economy. By introducing a robust foundation for agent-to-agent applications, the platform encourages the emergence of self-organizing, intelligent systems that can evolve over time. Its core framework is designed to be accessible to a wide range of developers, offering support for multiple programming languages—including Python and TypeScript—to foster innovation across disciplines. System architecture SpoonOS is structured across four layers: Data, Execution, Coordination, and Application. Its Core Framework is designed for developer accessibility and security. It integrates a Trusted Execution Environment for secure model inference and hosts a Model Context Protocol client to interface with various data sources, intended to encourage the development of new MCP servers. The framework includes BeVec, a proprietary vector database, and MCP+, an interface layer enabling seamless integration with on-chain and off-chain data. These components support agentic workflows by enabling access to diverse datasets while preserving privacy and security. Spoon core components (Source: Spoon) SpoonOS provides three primary capabilities: data availability and scalability, agent interoperability, and privacy and security. These components function together across the system’s layered architecture. Data availability and scalability MCP+ serves as a connection point between SpoonOS and decentralized storage, traditional databases, and processed Web3 datasets. It enables agents to access data across sources in a consistent format. BeVec, the system’s memory layer, supports semantic retrieval and retrieval-augmented generation from Web2 and Web3 sources. This design allows agents to retrieve contextual information efficiently and deliver informed outputs in real time. Agent interoperability Agents built on SpoonOS coordinate and share state using the AI Agent Interoperability Protocol. This protocol operates within the Coordination Layer and supports trustless collaboration through the use of Decentralized Identifiers and Zero-Knowledge Machine Learning. These technologies allow agents to verify results without exposing underlying data, enabling secure multi-agent task execution. Privacy and security The Privacy Module combines zero-knowledge proofs, fully homomorphic encryption, and Trusted Execution Environments to enable confidential and verifiable computation. A Security Monitoring component enforces output integrity through source verification, response cross-validation, and bias detection. This ensures agents operate within transparent and auditable parameters. Next steps To drive early adoption, SpoonOS is launching a US $2 million developer incentive program. It will including funding and support initiatives such as the Global Beacon Program, Global Developer Program, and Ecosystem Cooperation, with more details still to be announced. Hackathons and other events will also be used to incubate projects and explore use cases. Those interested in learning more about SpoonOS can visit the link below:https://spoonai.io/