Security at Clerxi.

Data protected in transit and at rest.

All network traffic between the Clerxi web platform, the Clerxi and Worksite Sentinel mobile apps, and our API at api.clerxi.com is encrypted using TLS 1.2 or higher. Application databases, object storage in S3, and rotating backups are encrypted using AES-256 via AWS-managed keys (AWS KMS). On mobile, authentication session tokens are stored in the device’s hardware-backed secure keystore — iOS Keychain on Apple devices and Android Keystore on Google devices — so a lost or stolen phone does not yield a usable session token to an attacker without device-level biometric or passcode unlock.

Account passwords are stored only as salted one-way hashes. Clerxi staff cannot read your password under any circumstances, and password material never leaves our servers in any reversible form.

Role-based access, scoped to your Organization.

Every record in Clerxi — projects, RFIs, daily reports, change orders, claim events, notice letters, certified payroll uploads, clock-in events — is bound to a single Organization tenant. Tenant isolation is enforced at the application layer on every read and write. Role-based access control (RBAC) determines which users in your Organization can view, edit, or approve which records. Roles and permissions are configured by your Organization administrator and can be tightened to the project level for sensitive jobs such as Department of Labor prevailing-wage audits or litigation-bound claim files.

Access to Clerxi production systems by engineering personnel requires multi-factor authentication, is restricted to a small named list, and is logged. Customer data is not accessed by Clerxi staff except where required to deliver support — and only on a documented request from the Organization administrator.

Append-only record of material actions.

Material actions on the Service are recorded in an append-only audit log that captures who acted, when, on what record, and what changed. The audit trail covers claim state transitions, notice generation and delivery, evidence linking, AI inference results (including which model produced the result and which user confirmed or rejected it), and administrative changes such as user invites, role changes, and integration credential updates. Audit logs are retained for at least 12 months for security and compliance purposes and are scoped to the Organization that owns the record.

Zero retention. No training on your data.

Clerxi’s AI-powered features — document analysis, claim classification, notice drafting, evidence linking, anomaly detection — rely on Anthropic’s Claude models served through the Anthropic API and through AWS Bedrock as an additional inference path. Both providers operate as processors under written Data Processing Addenda with Clerxi.

Anthropic does not retain prompts or responses beyond the duration of the API request under its zero-retention enterprise terms. Neither Anthropic nor AWS uses submitted content to train or improve foundation models. All AI-inference data is processed in US infrastructure. Clerxi itself does not train foundation models on customer data. Per-Organization operational baselines used for anomaly detection are scoped to a single tenant and never shared across customers.

SOC 2 Type II in progress.

We have implemented the technical and administrative controls required by the SOC 2 Type II Trust Service Criteria for Security, Availability, and Confidentiality. The third-party Type II audit is in progress. Until the audit report is issued, we do not represent ourselves as SOC 2 Type II certified.

Documented IR plan and honest disclosure.

Clerxi maintains a documented incident-response plan covering detection, triage, containment, eradication, recovery, and post-incident review. In the event of a data breach affecting personal information, we will notify affected customers and applicable regulators in accordance with the law that applies to the impacted data — for example, US state breach-notification statutes and, where relevant, GDPR Article 33 / 34 for EU/UK personal data. We will be specific about scope, impact, and remediation rather than vague.

Continuous scanning. Responsible disclosure welcome.

Our build pipeline runs automated dependency scanning against application and infrastructure code, flags security-relevant findings against severity thresholds, and gates production deploys on the resolution of high-severity issues. Clerxi also engages periodic third-party penetration testing against the platform; findings are triaged into the regular engineering backlog with priority weighting based on exploitability and blast radius.

If you discover a vulnerability in the Clerxi web platform, mobile apps, or API, please report it responsibly to support@clerxi.com with “Security report” in the subject line. We will acknowledge receipt promptly, keep you informed as we triage, and credit researchers who request acknowledgement once a fix is shipped.