Built for companies that take data seriously
Row-level isolation. No shared schemas, no shared queries.
Every query carries a tenant_id that Postgres applies via an automatic RLS policy. If an API forgets the predicate, the database adds it — and a regression test fails the build in CI.
- Single Postgres DB with RLS per table
- tenant_id set on auth, not in queries
- Automated tests verify `SET ROLE other_tenant` returns no rows
- Immutable per-tenant audit log — append-only on separate partition
Defense in depth — from the row to the cert
Multi-tenant with RLS
Every company has row-level isolation in Postgres. No query crosses the tenant perimeter — verified by automated tests that run on every deploy.
Encryption in transit and at rest
TLS 1.3 for all HTTP traffic. AES-256-GCM for data on disk with keys rotated quarterly. Backups encrypted with separate keys.
SSO + SCIM
Google Workspace, Microsoft 365, Okta, Azure AD. Automatic user provisioning on enterprise plans, with bi-directional sync.
Immutable audit logs
Every action that touches customer data is logged with actor, resource, IP, timestamp, and diff. Exportable to CSV or via signed webhook to your SIEM.
Data residency
US-East by default. EU-West for clients who require it (GDPR). No cross-region transfers without your explicit consent.
Compliance roadmap
SOC 2 Type II in progress (initial audit Q3 2026). HIPAA-ready for healthcare. ISO 27001 evaluation underway.
How each release gets built.
Mandatory code reviews
Every PR requires at least one senior engineer approval + explicit documented bypass for hotfixes. Zero force-push to main.
SAST on every PR
Semgrep + CodeQL run in GitHub Actions before merge. Custom rules for RLS anti-patterns, tenant_id leaks, and secrets in code.
Nightly DAST
OWASP ZAP scans staging nightly with different auth roles. High+ severity findings auto-open a ticket in < 1 hour.
Audited dependencies
Dependabot + Snyk monitor CVEs. Critical patches in < 24h, high in < 7 days. Zero deps without license review.
Zero secrets in repo
TruffleHog + GitHub secret scanning block pushes with tokens. All secrets in Vault with automatic quarterly rotation.
Safe releases
Canary deploys with feature flags + rollback in < 60s. Each release ships a signed changelog with SBOM (Software Bill of Materials).
Your data doesn't train public models.
We use foundation models through enterprise contracts (OpenAI, Anthropic, Google) with contractual guarantees: your data is not used to train public models. All prompts and responses live in your tenant, audit-logged and deletable on demand.
- No training on your data — contractually guaranteed
- Prompts and responses live in your tenant, never cross-tenant
- Deletable on demand via API or from the dashboard
- Optional auto-redaction of PII before LLM call
- Audit log of every Co-Pilot prompt and response
- Opt-out at workspace, team, or individual user level
Things you can do without asking support.
Export anytime
Button at /settings/data → ZIP with JSON + attachments of everything in the workspace. No tickets, no 30-day waits, no retention hostage.
Delete on demand
Hard delete in < 24h. Soft delete with 7-day recovery window for accidents. Backups purge in < 30 days.
Configurable retention
Per workspace: define how long logs, AI prompts, and files live. Default 365 days for logs, 90 for AI prompts. Your call.
Invite controls
Domain restriction, MFA required, invite expiration. Owner approves every new external domain.
How we test ourselves — and who else gets to test us.
Internal pentest
In-house security team + external consultant. Findings with severity > medium are published on this site in < 30 days.
Independent external pentest
Certified firm (CREST / OSCP). Public executive report; technical report under NDA for customers.
Bug bounty
Private HackerOne program with public scope. Bounties from $250 USD (low) to $10,000 USD (critical RCE / RLS bypass).
Threat modeling for critical changes
PRs touching auth, RLS, billing, or AI go through threat-model review before merge. Documented at /docs/security.
What we handle, what's yours.
Incident response
Documented runbook, 24/7 on-call with a 24-hour notification SLA, public post-mortems for impactful incidents, and a single email (incidents@florioin.com) that opens a ticket in our tracker.
Sub-processors
Public list maintained on this site: AWS (hosting), OpenAI/Anthropic/Google (Co-Pilot under enterprise contract), Resend (email), Cloudflare (CDN/WAF). 30-day notification before adding a new one.
Need SOC 2, a DPA, or a security questionnaire?
On-demand trust center — we send the latest report, fill questionnaires, and sign DPAs by email.
security@florioin.comReady to accelerate your business?
Request access today. We'll set you up within 24 hours.