Skip to main content
ForensicTrail
  • Platform
  • Sample case
  • For accountants
  • For attorneys
  • Trust
ContactClient login Book a demo
Menu
  • Platform
  • Sample case
  • For accountants
  • For attorneys
  • Trust
  • Contact
  • Client login
  • Book a demo

Trust

Subprocessors

The infrastructure, document-processing, and language-model providers that may process customer information to operate ForensicTrail.

Effective date
October 13, 2026
Version
2026-10-13

Launch review: Provider names and technical purposes are code-backed. Counsel must verify the executed DPAs, regions, retention, training, transfer mechanisms, and notice obligations before launch.

On this page

  1. Amazon Web Services
  2. Microsoft Azure Document Intelligence
  3. OpenAI
  4. Anthropic
  5. No silent assurance claims
  6. Changes

Amazon Web Services

Purpose: application hosting, private object storage, databases, queues, functions, identity integration, email, secrets, key management, logging, backups, and monitoring. Data can include account information, customer case data, evidence objects, derived work product, audit and custody records, identifiers, and operational metadata. Production resources are configured in the United States region selected by the deployment (currently us-east-1 in the deployment scripts).

Microsoft Azure Document Intelligence

Purpose: optical character recognition and document-layout extraction. Uploaded financial and case documents, or representations of their pages, may be sent for extraction. Contractual retention, training, and regional-processing commitments must be verified against the executed Microsoft agreement before launch.

OpenAI

Purpose: configured language-model features including the case assistant, producer authoring and reading, classification, paystub enrichment, credit-report extraction or review, asset-document extraction, and model-extraction fallback. The exact content varies from selected fields and transactions to document layouts, full statement content, or full case context. Requests are configured with store:false, but ForensicTrail does not currently claim a Zero Data Retention agreement or a verified universal no-training term. Those terms must be checked against the executed OpenAI agreement before launch.

Anthropic

Purpose: an available language-model provider for the same provider-neutral feature layer, plus a retained legacy Python assistant path in the PDF-service image. Which features use Anthropic depends on deployment configuration. Contractual retention and training terms must be checked against the executed Anthropic agreement before launch.

No silent assurance claims

This list identifies operational providers and purposes; it does not claim that any provider gives ForensicTrail a particular certification, Zero Data Retention, no-training commitment, geographic restriction, or customer-specific legal protection. Those commitments exist only if they appear in the executed customer and provider agreements.

Changes

The list is updated when a provider is added, removed, or given a materially different processing purpose. Customers requiring advance notice or an objection process should include those terms in their signed agreement. Questions may be submitted through the Contact page.

ForensicTrail

From financial records to the full money story.

Product

  • Platform
  • Sample case
  • Trust & methodology

Who it's for

  • Forensic accountants
  • Family-law attorneys

Company

  • Book a demo
  • Contact
  • Client login

Legal

  • Privacy Policy
  • Terms of Service
  • Subprocessors

© 2026 ForensicTrail. All rights reserved.