SECURITY
Mechanisms, not badges.
We would rather show you how the platform is built than hand you a logo. Everything below is shipped and in production today — described precisely enough that your security team can push back on it.
How isolation actually works
VizIQ is a multi-tenant SaaS. Every customer’s data lives in shared Postgres tables partitioned by workspace, not in a private database per customer. Isolation is enforced in two independent places: a cryptographic authentication gate in application code, and row-level security policies on every tenant table. Either one failing does not by itself expose another tenant’s catalog.
What is in place
The v1.5 platform hardening pass (July 2026) established the current posture. These are the controls, in the terms your reviewer will want them.
Authentication
The in-code gate is the only gate, so it is cryptographic.
- User tokens are verified locally against JWKS on every request — not trusted from the gateway.
- Service-role keys are compared in constant time, closing timing side channels.
- Inbound webhooks (Stripe, Telegram, store connectors) are signature-verified before any side effect, and mutations are idempotent on the provider event ID.
Access control
Who can see and change what, at both the app and database layer.
- Five workspace roles — owner, admin, creative ops, vendor and viewer — with permissions enforced per action.
- Row-level security is enabled on every tenant table, using membership-based policies rather than blanket allow rules.
- Privileged database functions run with a pinned search path, preventing schema-injection privilege escalation.
Data & money integrity
Encryption at rest, and no double-spends under concurrency.
- Third-party credentials (AI provider keys, bot tokens) are encrypted at rest with AES-256-GCM under a per-environment key.
- Ciphertext is stored in a versioned envelope, so keys can be rotated and rows migrated in place.
- Credit debits go through a single atomic operation that locks the row and is idempotent on a caller-supplied key — retries and concurrent dispatch cannot double-charge.
Network & abuse controls
What the platform is allowed to reach, and how often.
- An egress guard blocks server-side requests to private address ranges and URLs carrying embedded credentials.
- Rate limits apply to authenticated API surfaces, with configurable per-key ceilings.
- API keys are scoped, expiring, and revocable, with signed webhooks per store.
What we do not claim
A security page is only useful if the absences are as clear as the presences. So, plainly:
- — VizIQ does not hold SOC 2, ISO 27001 or equivalent third-party certification. If your process requires one, tell us early — it changes the timeline, and we would rather say so now than at contract stage.
- — We do not offer SAML or OIDC single sign-on today. Authentication is email and password, magic link, or Google.
- — We do not offer single-tenant or dedicated-database deployment. The isolation model is the shared-table model described above.
Security questionnaire, architecture review, or a question this page does not answer? Talk to us →
CASE STUDY · LIVE DEPLOYMENT
Thousands of products, every one of them hitting Super-Pharm's 6-image standard.
Super-Pharm runs a Mirakl-powered marketplace — and its providers use VizIQ to hit that marketplace’s visual standard. Both sides of the same catalog, live.
Read the full story →