How does Glen relate to ADK's session state and memory service?

Session state covers the current interaction; ADK's memory service is an interface you must back yourself. Glen is a ready, durable, org-shared memory you connect over MCP, persisting across sessions, deployments, and agents.

Can I use Glen with ADK agents on Vertex AI?

Yes. Glen is a standard remote MCP server reachable over OAuth or an API key, so ADK agents deployed anywhere, including Vertex AI, can read and write the same shared memory.

Shared memory for Google ADK

Google's Agent Development Kit gives you a structured way to build, evaluate, and deploy agents, with sessions, tools, and a path to running on Vertex AI. ADK has a notion of session state and even a memory service interface, but standing up a production memory backend, wiring it across multiple agents, and making sure what one agent learns is available to the others is work you own. By default each ADK agent's useful context is scoped to its session or its own service instance, so knowledge does not naturally flow between agents or survive a redeploy. Glen, shared memory for AI agents, gives your ADK agents one durable, org-wide memory exposed as a single MCP tool.

Expose Glen to your ADK agents over MCP and any agent can call one tool that both retrieves relevant long-term context and records new facts in a single round trip, complementing ADK's session state rather than replacing it. Session state handles the current interaction; Glen holds the durable knowledge that should persist long after the session closes and be shared with every other agent you deploy. An agent can read what the organization already knows before it acts, then write back what it learned, so you avoid hand-building a memory service, choosing a store, and maintaining a write path just to give your agents memory that lasts.

Because Glen is org-scoped, the memory layer spans every ADK agent, every deployment on Vertex AI or elsewhere, and every other agent in your organization, instead of being confined to one agent's memory service instance. One agent learns a durable fact; another ADK agent, or an agent on a completely different framework, reads it next time. That cross-agent, cross-deploy persistence is the part ADK leaves to your infrastructure. And because Glen is a standard MCP server, the same memory your ADK agents write is readable from Claude Code, Cursor, or any other MCP client, so your Google-built agents and the rest of your tooling share one knowledge base. Connect it once over OAuth or an API key and let the memory compound.

FAQ

How does Glen relate to ADK's session state and memory service?: Session state covers the current interaction; ADK's memory service is an interface you must back yourself. Glen is a ready, durable, org-shared memory you connect over MCP, persisting across sessions, deployments, and agents.
Can I use Glen with ADK agents on Vertex AI?: Yes. Glen is a standard remote MCP server reachable over OAuth or an API key, so ADK agents deployed anywhere, including Vertex AI, can read and write the same shared memory.