elizaOS Teams Up with Walrus to Power Agentic AI Memory for Decentralized Multi-Agent Workflows

elizaOS, the open-source platform for building and orchestrating autonomous AI agents, has integrated Walrus to provide a trustworthy foundation for agent memory, datasets, and cross-agent workflows. By embedding Walrus into the onboarding flow with time-limited credits, new users can immediately leverage the platform’s memory capabilities, reducing setup friction and accelerating adoption.

Walrus addresses key challenges in agentic AI by providing a unified data infrastructure for large models, datasets, and media. Every upload resolves to an on-chain proof-of-availability certificate on Sui, ensuring provenance and auditability. This enables developers to store, retrieve, and share data in a decentralized, verifiable way, making Walrus the foundational memory layer for multi-agent AI workflows.

“One of the biggest challenges in agentic AI is creating persistent, verifiable memory for agents that can be shared securely across workflows,” said Rebecca Simmonds, Managing Executive of the Walrus Foundation. “By integrating Walrus with elizaOS, we’re establishing a memory layer that makes it easy for developers to store, retrieve, and orchestrate data across agents while maintaining provenance, security, and auditability.”

This integration advances several goals. By making Walrus the default data platform in elizaOS V2, developers can immediately manage agent memory and data artifacts, experiencing the platform’s full memory capabilities from day one. Adoption is further encouraged through case studies, developer testimonials, and a portfolio of projects demonstrating real-world utility. The integration also simplifies onboarding with documentation, examples, and production-ready recipes, while showcasing Walrus’ multi-chain potential beyond the Sui ecosystem.

“Walrus provides the memory layer elizaOS needed to scale multi-agent systems efficiently and securely,” said Shaw Walters, Founder of elizaOS. “Developers can now focus on creating advanced agent workflows without worrying about fragmented data management, and this integration highlights the potential of decentralized, verifiable memory in real-world AI applications.”

The partnership also includes support for Seal and Nautilus, which enable encrypted access control and confidential off-chain computation. Together with Walrus, they allow developers to define how agents can access, process, and monetize data – forming the backbone of a secure, verifiable, and data-driven AI economy.

As elizaOS v2-powered agents and products launch later this year, Walrus is positioned to become the universal data layer for agentic AI, bridging storage, access control, and programmability into one trusted platform for the next generation of decentralized intelligence.

‍