Evidence Integrity

How PatternProof preserves your original records — and what our integrity checks actually prove.

Originals are preserved unchanged

When you upload a file, PatternProof stores the original bytes exactly as received. Previews, thumbnails, OCR text, transcripts, redacted copies, and export copies are stored separately as derivatives. Editing or redacting never overwrites the original.

SHA-256 hashes

Every preserved original is fingerprinted with SHA-256. Derivatives get their own hash. Your professional-review export includes a hash for every file plus a root hash of hashes.

A SHA-256 hash only proves that the stored bytes match the preserved version. It does not prove truth, authorship, creation date, or admissibility. Any court or professional judgment about a file is separate from this integrity check.

What we record with each file

Original filename, byte size, and MIME type
Upload timestamp and reported source
Raw metadata as received, and a normalized copy
Preservation status: received, preserved, extraction pending, needs attention, unsupported-but-preserved, or upload incomplete
Version history and duplicate relationships

Duplicates and corroboration

Duplicate files (renamed, re-encoded, forwarded, or imported through multiple sources) are grouped into an evidence family. PatternProof will show "N files represent 1 underlying record" rather than counting the same message six times as six pieces of proof.

What we do not claim

PatternProof does not prove abuse, determine legal truth, guarantee admissibility, or replace professional judgment. It preserves and organizes what you already have.

This page describes controls that are implemented in PatternProof today. If a description does not match your experience, please tell us. Related: Privacy · Safety · Evidence integrity · AI transparency · Professional access