Verdictums

Company / Security

Security & Defensibility

Last updated: September 2026

Verdictums exists because handling evidence demands a higher standard than general-purpose AI was built to meet. For us, security and defensibility are the same problem. This page summarizes how the engine is designed and operated so that partners can rely on its outputs when they reach a hearing. We are an early-stage company and will update this page as our controls and certifications mature.

01Deterministic processing

Every deployment runs pinned model versions. Given identical input, the engine produces byte-identical output that can be reproduced on demand, months or years later. Reproducibility proofs are a standard part of what a pilot delivers.

02Tamper-evident chain of custody

Each file is hashed on intake, and every later operation, whether processing, redaction, or export, is appended to a hash-chained audit log. If a file or a log entry is altered, the chain breaks and the alteration is detectable.

03Built-in redaction

Faces, license plates, and audio segments are redacted within the pipeline itself rather than as a post-processing step. Redaction events are recorded in the audit log with their provenance, and redactions already present in source documents are detected and flagged.

04Data isolation and encryption

Partner evidence is logically isolated per tenant and encrypted in transit and at rest. We do not use partner evidence to train models. Access is limited to what is needed to operate the Service under the partner's instructions, and retention follows the partner agreement.

05Access control and monitoring

API access is authenticated and scoped per partner. Internal access follows least-privilege principles, and administrative actions are logged and reviewable, which is the same discipline the engine applies to evidence.

06Infrastructure

The Service runs on established cloud infrastructure with regional data residency options agreed per partner. Infrastructure details, sub-processors, and any relevant certifications are provided to partners during due diligence under NDA.

07Responsible disclosure

We welcome reports from security researchers. If you believe you have found a vulnerability in the Site or the Service, please write to contact@verdictums.com with the details. We will acknowledge your report and keep you informed while we investigate.