Identity-bound access
Cohort invitations are tied to the intended sign-in email, expire automatically and cannot be reused.
- Server-side membership checks
- Owner, instructor and learner roles
- Private personal chats
Security designed into the learning loop
Figurve protects learning data in transit and at rest, and treats identity, course sources and model context as separate trust boundaries.
Controlled source ingestion
Uploads are type, signature and size checked before storage. URLs use a restricted importer and extracted passages retain provenance.
- Private object storage
- Exact passage citations
- Per-workspace source scopes
Encrypted by default
Data is encrypted in transit and at rest within Figurve-managed systems. Access is limited by server-side authorization and private credentials never ship to the browser.
- TLS-protected connections
- Encryption at rest
- Server-held API credentials
Bounded provider processing
Source text is treated as untrusted reference content. When you enable a model, transcription or live web-search feature, its authorized provider processes only the request data needed for that feature. The provider's API data controls and retention terms may apply.
- Prompt-injection boundary
- Explicit source controls
- No browser-supplied roles or user IDs
Your data stays yours
You retain ownership of your uploads and learning work. Figurve never uses that content to train AI models, and account controls let you export or delete your data. Shared cohort submissions may follow the cohort's stated retention policy.