How is this different from a data catalog or dbt docs?

Catalogs and docs are references humans must consult. Glen gives agents memory they retrieve and write automatically in one MCP call, so the right definition and the relevant caveat surface while the agent is actually writing the query or report.

Will every analyst's agent use the same definitions?

Yes. Glen is org-shared, so every analyst's agents read and write the same store. A metric definition or data caveat one person records is immediately applied by everyone else's agents, keeping numbers consistent.

Shared memory for analysts

Analysts spend half their time rediscovering context: which table is the source of truth, how revenue is actually defined this quarter, why a metric jumped in March, the join that everyone gets wrong. That knowledge lives in tribal lore, in dashboard footnotes, and in the analyst who built the original model, so every new question starts with archaeology. When analysts adopt AI agents to write SQL, build reports, or interpret results, the agents have no idea about your schema's quirks or your business definitions and confidently produce numbers that are subtly wrong. Glen, shared memory for AI agents, gives those agents a durable, team-shared memory through a single MCP tool that retrieves the relevant analytical context and records new definitions and findings in one round trip.

Point the agents your analytics team uses at Glen over MCP and each one gains a single tool that reads the relevant context and writes back what it learns. Before an agent writes a query, it pulls what the team already knows: that active users excludes internal accounts, that the canonical orders table is the deduped one, that a spike in a metric was a tracking bug, not a real change. After it answers a question, it records the definition it used and the caveat it found, so the next agent, and the next analyst, inherit a consistent understanding instead of reinventing it. You stop re-explaining the data model and stop shipping reports that quietly contradict each other.

The core change is that analytical context becomes shared and durable rather than locked in one analyst's head. Glen is org-scoped, so the memory spans every analyst and every agent on the team, and even stakeholders' agents that ask the same questions. One analyst's agent records the correct revenue definition; every other agent uses it, so numbers reconcile across reports. Because Glen is a standard MCP server, the same memory is readable from any MCP client, so a SQL-writing agent, a reporting assistant, and an ad-hoc question bot all draw on one consistent record of how your data and your business actually work. Connect once over OAuth or an API key and let your team's analytical knowledge compound instead of resetting with every new hire and every new question.

FAQ

How is this different from a data catalog or dbt docs?: Catalogs and docs are references humans must consult. Glen gives agents memory they retrieve and write automatically in one MCP call, so the right definition and the relevant caveat surface while the agent is actually writing the query or report.
Will every analyst's agent use the same definitions?: Yes. Glen is org-shared, so every analyst's agents read and write the same store. A metric definition or data caveat one person records is immediately applied by everyone else's agents, keeping numbers consistent.