We'd rather a security review start from an accurate baseline than a marketing claim. Here's exactly what's enforced today and what's actively being built.
Every table carrying tenant data is scoped by tenant id, and every query filters on it. There is no cross-tenant query path in the product's API layer.
Every organization gets its own login and subdomain, with admin/viewer roles enforced on every write and admin-only route.
Every API request is logged — masked actor identity, role, tenant, method, path, status, and latency — independent of the request path itself.
All credentials (AI provider keys, session signing key, database URL) are environment-driven — nothing is committed to source.
Customer identifiers used across the product's demo data are opaque, hashed tokens — never real names, emails, or account numbers.
Login sessions are short-lived, signed tokens — not indefinite credentials — reducing the blast radius of a leaked session.
These are in progress, not implied as done. Ask your sales contact for current timelines if any of these gate your evaluation.
| Area | Status | Detail |
|---|---|---|
| Encryption at rest | Roadmap | Planned before any deployment stores real (non-synthetic) customer transcripts. |
| Enterprise SSO (SAML/OIDC) | Roadmap | Org-based email + password login is live today; federated SSO is planned for larger deployments. |
| Retention enforcement | Roadmap | Retention windows are configurable per plan today, but not yet enforced by an automated purge job. |
| Field-level access control | Roadmap | Beyond today's admin/viewer roles, for fine-grained access to sensitive transcript fields. |
| SOC 2 | Roadmap | Not yet initiated. If this gates your procurement, tell us — it shapes our timeline. |
We'll answer it directly, without overstating what's built.