How is Glen different from Bedrock session memory and knowledge bases?

Bedrock session memory keeps one conversation coherent and knowledge bases retrieve from documents you ingest. Glen is durable, org-shared long-term memory that every agent writes to and reads from, combining retrieval and writing in one MCP call.

How does a Bedrock Agent connect to Glen?

Glen is a standard MCP server. Bedrock Agents that support MCP tools can call Glen directly; otherwise expose it through an action group or gateway that proxies the MCP call to retrieve context and record observations.

Shared memory for Amazon Bedrock Agents

Amazon Bedrock Agents let you build managed agents on AWS that plan multi-step tasks, call action groups, and ground responses in knowledge bases backed by your own data. Bedrock added session memory so an agent can retain context across turns within a conversation, and knowledge bases handle retrieval over documents you ingest. What sits between those two, durable facts an agent should accumulate across sessions and share with every other agent in your organization, is still something you assemble yourself out of DynamoDB tables, Lambda glue, and bespoke retrieval. Glen, shared memory for AI agents, gives Bedrock Agents that long-term shared memory as a single MCP tool that both reads context and writes new facts in one round trip.

Connect Glen to a Bedrock Agent over MCP and the agent can call one tool that retrieves relevant long-term organizational context and records what it learned during the task. At the start of a task it pulls what the org already knows about an account, a customer, or a system; as it completes action groups it writes back the durable facts it discovered. You stop hand-building memory infrastructure, the DynamoDB schema, the Lambda write path, the embedding job, just to give a Bedrock Agent knowledge that persists beyond a session and is shared with the rest of your fleet.

Bedrock session memory keeps a single conversation coherent, and knowledge bases serve static documents you ingest; neither is a living, org-shared fact base that every agent updates. Glen is org-scoped, so its memory spans every Bedrock Agent, every AWS account boundary you connect, and every other agent your organization runs in any framework. One agent learns something durable and another reads it next time without you replicating state across services. Because Glen is a standard MCP server, the same memory your Bedrock Agents write is readable from Claude Code, Cursor, or any MCP client, so human and agent work share one knowledge layer. Wire it in once over OAuth or an API key and the memory compounds across sessions and agents.

FAQ

How is Glen different from Bedrock session memory and knowledge bases?: Bedrock session memory keeps one conversation coherent and knowledge bases retrieve from documents you ingest. Glen is durable, org-shared long-term memory that every agent writes to and reads from, combining retrieval and writing in one MCP call.
How does a Bedrock Agent connect to Glen?: Glen is a standard MCP server. Bedrock Agents that support MCP tools can call Glen directly; otherwise expose it through an action group or gateway that proxies the MCP call to retrieve context and record observations.