How is this different from LangGraph's checkpointing and state?

LangGraph state and checkpoints persist a single run's progress. Glen is durable, org-shared long-term memory that outlives any run and is read and written by every agent in your organization.

How do my LangGraph agents reach Glen?

Glen is a standard MCP server. Connect to it as an MCP tool, then call it from any node to retrieve context and record observations in one round trip.

Shared memory for LangGraph

LangGraph gives you fine-grained control over agent workflows, nodes, edges, and state machines that route work through tools and decisions. It has solid primitives for short-term state and checkpointing within a run, but cross-run, cross-agent knowledge, the durable facts your graphs should accumulate over weeks and across deployments, is something you end up building and babysitting yourself. Glen, shared memory for AI agents, gives your LangGraph agents that long-term shared memory as a single MCP tool, so every graph in your organization reads from and writes to one store instead of each maintaining its own.

Expose Glen to your LangGraph agents over MCP and any node can call one tool that both retrieves relevant long-term context and records new facts in a single round trip. A retrieval node can pull what the organization already knows about an entity or task before the graph decides what to do; a later node can write back what the run discovered. You stop hand-rolling a vector store, an embedding pipeline, and a write path just to give your graph durable memory, and you stop wiring per-agent silos that never share what they learn.

Because Glen is org-scoped, the memory layer is shared across every graph, every deployment, and every other agent in your org rather than bound to a single LangGraph process. One workflow learns a durable fact about a customer or a system; a different workflow, even a different framework, reads it next time. That is the part LangGraph's built-in state does not give you: persistence that outlives the run and is shared across agents. And because Glen is a standard MCP server, the same memory your LangGraph agents write is readable from Claude Code, Cursor, or any other MCP client, so human-driven and graph-driven work draw on the same knowledge. You wire it in once over OAuth or an API key and let the memory compound.

FAQ

How is this different from LangGraph's checkpointing and state?: LangGraph state and checkpoints persist a single run's progress. Glen is durable, org-shared long-term memory that outlives any run and is read and written by every agent in your organization.
How do my LangGraph agents reach Glen?: Glen is a standard MCP server. Connect to it as an MCP tool, then call it from any node to retrieve context and record observations in one round trip.