Most multi-tenant breaches are a missing filter in one query. We assume that will eventually happen and put the boundary somewhere application code cannot bypass.
Every request resolves an organization context before any handler runs, and a database-layer extension independently injects that scope into queries across every multi-tenant model. A missing check in application code cannot leak another tenant's data, because the second layer does not depend on the first.
SAML and OIDC single sign-on against your identity provider, SCIM provisioning and deprovisioning, TOTP multi-factor with backup codes, trusted-device management, and short-lived access tokens with rotating refresh tokens.
Role-based access control with per-organization roles and granular resource.action permissions. Every protected route composes authentication, membership resolution and tenant enforcement — including AI agents, which can never exceed the permissions of the user they run for.
TLS in transit. Integration credentials, OAuth tokens, API keys and webhook secrets are encrypted at rest with a dedicated key, with support for key rotation via a previous-key window so re-encryption needs no downtime.
Security-relevant actions are recorded with actor, organization, resource and request id — including actions taken by agents. Logs are exportable, and correlate to application logs through a shared request identifier.
Configurable retention policies per data class, legal hold that suspends deletion for named records, and consent records tracked as first-class data with a PII registry classifying sensitive fields.
Any feature that fetches a user-supplied URL — workflow steps, webhooks, agent tools — resolves DNS and blocks private and link-local ranges before connecting, including on redirects, so internal services and cloud metadata endpoints stay unreachable.
Dependency audits block the build on any high-severity advisory. The metrics endpoint is authenticated. Secrets are supplied by environment and never committed, and the application refuses to start in production if required security configuration is missing.
Customer data is never used to train foundation models and is never shared between tenants. You can bring your own AI provider keys, and export all of your data in open formats at any time.
Voltx is in limited beta and does not yet hold a SOC 2 Type II or ISO 27001 certification — we would rather tell you that than imply otherwise. The controls above are implemented today and we are happy to walk a security reviewer through any of them, answer a vendor questionnaire, or discuss self-hosted deployment if your requirements need it now.
Found something? Email support@usevoltx.com and we will acknowledge within one business day. Please give us a reasonable window to fix an issue before disclosing it publicly.
Start free in minutes, or talk to our team about your rollout.