Database

MongoDB with Mongoose 9. Single shared database, multi-tenant via organizationId. 9 collections.

erDiagram
  Organization ||--o{ User : organizationId
  Organization ||--o{ Position : organizationId
  Organization ||--o{ Candidate : organizationId
  Organization ||--o{ Interview : organizationId
  Organization ||--o{ InterviewReport : organizationId
  Organization ||--o{ CreditLot : organizationId
  Organization ||--o{ CreditLedger : organizationId
  Organization ||--o{ Note : organizationId
  Position ||--o{ Interview : positionId
  Candidate ||--o{ Interview : candidateId
  Interview ||--|| InterviewReport : interviewId
  CreditLedger ||--o{ CreditLot : purchaseLedgerEntryId
  CreditLot ||--o{ CreditLedger : lotId
  Candidate ||--o{ Note : target
  Interview ||--o{ Note : target

Every tenant-scoped document carries organizationId. The API enforces isolation in queries; a few global unique keys exist for login and join links (users.email, organizations.slug, interviews.joinTokenHash).

Collections

Collection	Purpose
organizations	Tenant root — settings, branding, cached credit balances
users	Team members, roles, invites, Firebase auth linkage
positions	Job templates — skills, AI context, duration
candidates	People records, resume metadata, recruiter profile fields
interviews	Scheduled sessions — frozen snapshot, room/runtime, credits
interviewreports	Post-interview AI report — scores, transcript (1:1 with interview)
creditlots	Purchasable or granted credit pools with expiry
creditledgers	Append-only credit movements and balance snapshots
notes	Recruiter notes on a candidate or interview

View complete schema → DBML