Docs › Getting started › Concepts: deployments, tenants, memory
Concepts: deployments, tenants, memory
Last updated July 2026 · ~3 min read
Three concepts describe how MeetCrew is organized: the tenant, the deployment, and memory. Together they define who owns what, which teammate is in the room, and what that teammate remembers.
Tenant
A tenant is your organization’s isolated space in MeetCrew. In v1 one tenant maps to one Microsoft Entra tenant, and identity flows from Entra. Each tenant is provisioned with its own Azure resources for structured memory, search, and transcripts. Isolation between tenants is architectural, not a policy filter: one tenant’s services hold credentials only for that tenant’s resources, so there is no path from one customer’s data to another’s.
Deployment
A deployment is one named teammate — Jim, Todd, or Ariel. Each deployment has its own name (the word people use to address it), character, voice, standing instructions, and a bound backend agent that supplies its reasoning. A tenant can run several deployments at once, each configured for a different job: Jim for renewals, Todd for implementations, Ariel for legal review.
Memory
Memory is a structured, searchable record of the meetings a deployment attended: decisions, open questions, action items, statements, and topics, with owners and dates. Memory is scoped per deployment. Jim’s memory is private to Jim by default. You can configure a deployment to share its memory with a sibling, and sharing is asymmetric — granting Todd read access to Jim does not grant Jim access to Todd.
Cross-meeting recognition uses Microsoft Entra platform identity, never voice. MeetCrew stores no voiceprints, and anonymous speakers are a first-class part of the record.
Was this helpful? Yes · No | Still stuck? Contact support.