Last updated June 2026
Security
Your network data is critical infrastructure. OmniTwin treats every byte of it with the same rigor you would expect from the systems it manages.
Infrastructure Security
OmniTwin is hosted on cloud infrastructure from SOC 2 Type II certified providers. Backend services are intended to be reachable only from inside the deployment network, and network exposure is being audited and tightened ahead of general availability. We would rather say that than claim a posture we have not finished verifying.
- Production workloads run in containers, deployed from version-controlled images built by the pipeline.
- Infrastructure is provisioned and managed exclusively through version-controlled Infrastructure as Code (IaC), ensuring every change is auditable and reproducible.
- Infrastructure changes are made through the pipeline rather than by hand, so they are reviewable. Continuous configuration drift detection with alerting is being built out, and is not something we claim today.
- The deploy pipeline verifies database role grants before cutover and rolls back automatically if a step fails, so a bad deploy does not sit in production waiting to be noticed.
Data Encryption
All data is encrypted both in transit and at rest. We do not support unencrypted connections to any OmniTwin service.
- In transit: All connections use TLS 1.3 with strong cipher suites. We enforce HSTS with a minimum max-age of one year, including subdomains. Certificate transparency logging is enabled for all domains.
- At rest: All data stored in PostgreSQL and the graph database is encrypted using AES-256. Encryption keys are managed through a dedicated Key Management Service (KMS) with automatic rotation on a 90-day cycle.
- Backups: Database backups are encrypted with separate keys from production data and stored in geographically redundant locations with strict access controls.
- Secrets management: API keys, database credentials, and service tokens are stored in a hardened vault with audit logging on every access. Secrets are never stored in source code, environment variables, or container images.
Access Control
OmniTwin enforces the principle of least privilege at every layer of the stack, from application-level RBAC to infrastructure access.
- Authentication: All user accounts require strong passwords and support multi-factor authentication (MFA). SAML 2.0, OIDC, and automated Okta and Azure AD group mapping are included on standard tiers, not gated behind an enterprise contract. There is no SSO tax.
- Role-based access control: Granular RBAC allows organizations to assign specific permissions by role (viewer, operator, admin, owner) with full audit trails on all permission changes.
- API authentication: All API access requires scoped tokens with configurable expiration. Token usage is logged, and a revoked token stops working without a restart or a deploy.
- Infrastructure access: Production environment access is limited to authorized engineering staff through short-lived, just-in-time credentials. All production access is logged and reviewed weekly.
- Machine authentication: CI/CD and Terraform pipelines authenticate with in-memory OAuth 2.0 bearer tokens, so build scripts never carry hardcoded passwords.
- Tenant isolation: Multi-tenant data is isolated by PostgreSQL Row-Level Security, bound to the authenticated session and enforced by the database itself rather than by application code, under a role that cannot bypass it.
For a security reviewer, this is what a single request must clear before it can read or write tenant data, and what is refused along the way.
How one request is kept from other tenants
Five stages, all wired. The last one is the guarantee: the database refuses cross-tenant rows.
- Wired
Exactly one credential is accepted: an API key, a bearer token, or a JWT. A request carrying more than one is refused.
- Wired
The required scope for the method is checked. API keys are rate limited.
- Wired
The verified tenant is taken from the credential and stamped on the request. If the request body also names a tenant, it must match. The credential is the authority; the payload never is.
- Wired
Identifiers are checked before any query runs.
- Wired
The query runs under PostgreSQL row-level security, as a role that cannot bypass it. Isolation is enforced by the database, not by application code remembering a WHERE clause.
Network Security
Our network architecture follows a zero-trust model. No service implicitly trusts any other service, regardless of network location.
- Zero-ingress architecture: Backend processing nodes sit behind zero-ingress Cloudflare Tunnels. Public inbound ports are dropped at the OS firewall layer, so public internet traffic cannot be routed directly to database ports.
- All inter-service communication uses mutual TLS (mTLS) with certificate-based authentication.
- Network segmentation isolates production, staging, and development environments with no cross-environment connectivity.
- Web Application Firewall (WAF) rules protect all public-facing endpoints against OWASP Top 10 threats, including SQL injection, XSS, and CSRF attacks.
- DDoS protection is provided at the network edge with automatic traffic scrubbing and rate limiting.
- All network traffic is logged and analyzed for anomalous patterns using automated threat detection.
Application Security
Security is built into every phase of our software development lifecycle, from design through deployment.
- Secure SDLC: All code changes go through mandatory peer review with security-focused checklists. Automated static analysis (SAST) and dependency scanning run on every pull request.
- Dependency management: We continuously monitor all third-party dependencies for known vulnerabilities using automated scanning tools. Critical vulnerabilities are patched within 24 hours of disclosure.
- Container security: All container images are built from minimal base images, scanned for vulnerabilities before deployment, and run with read-only root filesystems and dropped capabilities.
- Input validation: All user input is validated and sanitized at the application boundary. Our Rust-based math engine provides deterministic, memory-safe CIDR validation that eliminates entire classes of overflow and injection vulnerabilities.
Incident Response
We maintain a documented incident response plan that is tested and updated quarterly.
- Detection: Automated monitoring and alerting covers all production services with sub-minute detection times for availability and security events.
- Escalation: Security incidents are classified by severity (P0 through P3) with defined response times: P0 critical incidents receive immediate response with customer notification within 1 hour.
- Communication: Affected customers are notified through multiple channels (email, in-app notification, status page) with regular updates until resolution.
- Post-incident review: Every incident triggers a blameless post-mortem within 48 hours. Findings and remediation actions are tracked to completion and shared with customers when relevant.
Penetration Testing
OmniTwin conducts regular security assessments to identify and remediate vulnerabilities before they can be exploited.
- Third-party penetration tests are conducted at least annually by qualified, independent security firms.
- Automated vulnerability scanning runs continuously against all production services and infrastructure.
- Results from all security assessments are tracked through remediation, with critical findings addressed within defined SLAs.
- Summary reports from third-party assessments are available to enterprise customers under NDA upon request.
Employee Security
Our team follows strict security practices to protect customer data and infrastructure access.
- All employees undergo background checks prior to employment and receive security awareness training during onboarding and annually thereafter.
- Endpoint protection is mandatory on all company devices, including full-disk encryption, screen lock policies, and endpoint detection and response (EDR) software.
- Access to customer data and production systems is granted on a need-to-know basis and reviewed quarterly.
- All employees sign confidentiality agreements covering customer data, proprietary technology, and internal security practices.
Vulnerability Disclosure
We welcome responsible disclosure of security vulnerabilities. If you discover a potential security issue in OmniTwin, please report it to security@kortesalabs.com.
- We will acknowledge receipt of your report within 24 hours.
- We will provide an initial assessment and expected timeline within 72 hours.
- We ask that you do not publicly disclose the vulnerability until we have had a reasonable opportunity to address it.
- We will not take legal action against researchers who follow responsible disclosure practices.