SOC 2
TYPE II
TYPE II
SOC 2
Type II
audited annually · since 2024
Independent attestation covering security, availability, confidentiality. Latest report covers 2025-Q3 → 2026-Q2.
Request report →QRM is single-tenant where it counts. No cross-tenant indexing. No training on your data. Every QaiS action is audited and reversible. Here's exactly how — in plain language.
Every workspace gets its own database row, blob namespace, and identity scope. Cross-tenant queries are blocked at the database, not just the application.
QaiS never trains on your tenant data. Prompts and context stay scoped to the workspace and are not retained by the model provider beyond the request.
Every QaiS tool call and every user action is logged with the actor, before/after diff, and one-click undo. Audit log retention is plan-tiered.
Full export anytime in JSON + raw blobs. 30-day backup window after cancellation, then a verified hard-delete with a signed certificate.
QRM is multi-tenant under the hood for cost efficiency, and single-tenant where it counts — at the database row, the file namespace, and the search index. Enterprise plans get full single-tenant deployments: their own database cluster, their own storage bucket, their own region.
Cross-tenant queries are impossible by construction. Tenant ID is enforced at the SQL row-security layer, not just the application code. Every read and write carries the calling tenant's ID, and the database refuses queries that try to read another tenant's rows.
You can run on a regional cloud of your choice: US-East, US-West, EU-West (Frankfurt), UK (London), or APAC (Sydney) on Enterprise.
QaiS uses third-party LLM providers (Anthropic, OpenAI) under data-processing agreements that prohibit model training on your data. Prompts and context are scoped to the calling tenant, sent over TLS, and the provider keeps them only for the duration of the request.
For Enterprise tenants, we offer self-hosted inference: QaiS runs against open-weight models inside your dedicated tenancy, so your data never leaves your region.
Every QaiS tool call is recorded with the user, the prompt, the tool args, the result, and a one-click revert. You can see the trace for any action your team or QaiS has ever taken.
Independent audits, on a schedule. We publish reports under NDA — email security@qais.app for a copy.
Independent attestation covering security, availability, confidentiality. Latest report covers 2025-Q3 → 2026-Q2.
Request report →Standard Contractual Clauses, EU data residency (Frankfurt), processor obligations met. DPA on Business and Enterprise.
View DPA →Information security management system. Pre-audit gap analysis complete. Certification expected Q1 2027.
Status →A complete, current list. We give 30 days' notice before any change on Business and Enterprise plans.
We page on real signals — not vague alerts. On-call engineers acknowledge within 5 minutes. We publish a public status page incident, send affected workspaces an email, and follow up with a post-mortem within 5 business days.
For security incidents that involve customer data, we notify affected workspaces within 24 hours of confirmation, regardless of contractual SLA. We retain transparency over face-saving every time.
Questions about controls, requests for audit reports, vulnerability disclosures, or DPAs — email security@qais.app. We respond within one business day; security disclosures get same-day.
Five minutes to open a workspace. A signed MSA on Enterprise.