Can the whole team read and write the same memory?

Yes. Memory stores belong to the organization and every member can read and write every store the org owns; provenance records who contributed each fact without limiting access.

Does it work across different editors?

Yes. Glen is a single MCP tool, so engineers on Claude Code, Cursor, or VS Code all share the same org-scoped memory.

Shared memory for engineering teams

On an engineering team, the same knowledge gets rediscovered over and over: how a subsystem fits together, why a past decision went the way it did, which sharp edges to avoid. As teams adopt AI coding agents, that waste multiplies, because every engineer's agent starts cold and learns the codebase independently, sharing nothing. Glen, shared memory for AI agents, is designed for this scale: one organization-wide memory that all of a team's agents read from and write to, delivered as a single MCP tool that works across Claude Code, Cursor, and VS Code.

The core idea is that knowledge should be a team asset, not a per-person one. When any engineer's agent learns something durable, an architectural constraint, a deployment quirk, a naming convention, a service that must never be called from a certain context, Glen captures it as an observation and links it into the org's shared memory. The next time anyone on the team touches related code, that knowledge surfaces automatically. Reads and writes happen in the same MCP round trip, so the team is not maintaining a wiki on the side; the memory accrues as a byproduct of the work agents are already doing.

For a team specifically, the org-shared scoping is the whole point. Memory belongs to the organization, every member can read and write every store the org owns, and provenance is tracked so you can see which engineer's agent contributed a given fact without that being an access boundary. New teammates inherit the org's existing memory the moment they join, with no per-user store to provision, so onboarding gets faster as the memory deepens. And because the same shared memory is available regardless of which MCP client an engineer prefers, the team gets one consistent base of knowledge across a heterogeneous toolset. The net effect is that your agents stop repeating each other's discoveries and start building on them, which is the compounding return engineering teams want from going agentic.

FAQ

Can the whole team read and write the same memory?: Yes. Memory stores belong to the organization and every member can read and write every store the org owns; provenance records who contributed each fact without limiting access.
Does it work across different editors?: Yes. Glen is a single MCP tool, so engineers on Claude Code, Cursor, or VS Code all share the same org-scoped memory.