Walrus Memory: Portable Agent Memory Live
Summary
The Upside announces Walrus Memory, a portable memory layer for AI agents. The post covers the new name and design that moves memory across apps and sessions, encrypted-by-default controls, shared multi-agent spaces, Python and TypeScript SDKs, and first-party plugins for OpenClaw and NemoClaw.
Walrus Memory is live 
New name, new look, more to ship.
If you've ever had an AI agent lose context, restart a workflow, or forget prior work, you've experienced the memory problem.
Walrus Memory gives agents portable memory so they can carry context across apps and sessions.
Portable by design
Memory moves freely across sessions and apps. No lock-in to a runtime or provider.
Fully under your control
Encrypted by default, with programmable access controls. Delegate or revoke at any time.
Built for coordination
Shared memory spaces keep multi-agent workflows aligned, with verifiable integrity built in.
Plugs into your stack
SDKs in Python and TypeScript
Native MCP support
First-party plugins for OpenClaw and NemoClaw
Out-of-the-box support for Claude, ChatGPT, and Gemini
Portable Memory Layer for AI Agents | Walrus Memory
Persistent agent memory that moves across apps and sessions. Walrus Memory enables reliable coordination and programmable access control for AI agents.
