ReportMax is built for law enforcement, engineered with CJIS Security Policy alignment from the ground up. The following summarizes our public-facing security architecture. For agency-specific security documentation including our full CJIS Security Plan, contact us directly.

Tokenization Architecture

Subject names and other direct identifiers are replaced with placeholder tokens on your own device before any information is transmitted for AI processing. Our AI subprocessor never receives raw subject-identifying information.

When you type a report, your browser automatically detects and replaces PII — full names, street addresses, dates of birth, phone numbers, and license plate numbers — with opaque tokens (e.g., [NAME_1], [ADDR_1]). Only tokenized text leaves your device. The mapping between tokens and real values exists only in browser memory and is discarded when the session ends.

Encryption

Replit’s published security documentation states that its PostgreSQL database service encrypts database traffic in transit using TLS and database storage at rest using AES-256 server-side encryption. ReportMax does not make broader public claims about the transport protocol used by every subprocessor connection or about the database service’s underlying key custodian.

Access Control

Audit Logging & Accountability

All authentication events, report access, edits, workflow transitions, AI generation requests, and exports are logged to a tamper-evident, cryptographically hash-chained audit trail. Each entry in the audit log is linked to the preceding entry via SHA-256 hash, making retroactive modification detectable.

Officers must certify review and accuracy before copying or exporting any AI-generated report. That certification is captured with officer identity and timestamp in the audit log and cannot be removed.

Data Retention

Draft report content is retained ephemerally and auto-deleted within 24 hours. ReportMax does not store finished report text on its servers. Officers copy finished reports to their agency's own Records Management System (RMS). Audit log metadata is retained for accountability purposes per law enforcement standards.

AI Provider Data Handling

ReportMax does not use submitted report content to train, fine-tune, or improve an AI model. Anthropic states that inputs and outputs from its commercial API are not used to train its generative models unless the customer opts in. Anthropic’s standard commercial API retention policy may apply to tokenized content sent for drafting; ReportMax does not currently claim that its Anthropic organization has Zero Data Retention.

Subprocessors

Subprocessor Function CJI Exposure
Anthropic (Claude API) AI narrative generation None — tokenized data only
Replit Hosting & database Account data, audit metadata, and temporary tokenized drafts
Stripe Billing None

For agency-specific security documentation including our full CJIS Security Plan, contact us directly.

Contact Us for Full Security Documentation