ENAnalizar tu unidad

Evidence standard

How ClearCopies decides what is exact—and what still needs you.

The scanner separates content evidence from keeper decisions. A match can show that bytes are identical; it cannot decide which path, share link, owner or version history matters to your work.

Last materially reviewed: August 19, 2026

Technical review · August 2026

The published rules match the running duplicate engine

Provider fingerprints are compared only within the same algorithm and byte count. Matching thumbnails can create a probable-media review group, but cannot become exact evidence when content fingerprints differ.

Exact
size + fingerprint
Probable
thumbnail + type + size
Protected
never auto-selected

The methodology was revised after comparing large-drive scan results with the review UI, including manual selections and protected formats.

Written by ClearCopies Editorial · Technical review by ClearCopies Engineering

Inventory

Enumerate provider metadata without downloading every original file.

Compare

Pair byte counts with provider-specific fingerprints.

Revalidate

Check every selected target again before a cleanup action.

Exact means size plus a comparable fingerprint

Equal names or sizes are useful ways to narrow a search, but they do not prove identical content. ClearCopies compares the byte count and a fingerprint supplied by the same cloud provider. Algorithms are stored with their values, so unrelated hash formats are never mixed.

OneDrive

quickXorHash, with SHA-1 when available

Microsoft Graph returns file hashes as read-only metadata. ClearCopies retains the algorithm name and compares only compatible values.

Microsoft Graph hashes resource

Google Drive

SHA-256, then SHA-1, then MD5 when available

Google-native Docs and shortcuts may not expose a binary checksum. Those items cannot become exact groups without comparable content evidence.

Google Drive files resource

Dropbox

Dropbox content_hash

Dropbox content_hash is a provider-defined block hash. ClearCopies does not relabel it as ordinary whole-file SHA-256.

Dropbox content hash reference

Four classifications, four different safety boundaries

Exact

Same byte count and the same comparable content fingerprint.

Eligible for a suggested removal plan unless a protected file type or another safety rule applies.

Probable media

Same file type, byte count and thumbnail fingerprint, but no matching content fingerprint.

Never treated as exact. The user must review and select copies manually.

Needs review

Mixed or incomplete signals that do not prove identical content.

No automatic removal recommendation.

Protected

A group includes formats such as CAD, project, design or database files that need additional context.

Kept outside automatic removal even when the bytes match.

Cleanup is a separate decision

  1. The user chooses which copies to keep and which to include in the plan.
  2. The checkout total reflects every selected file, including manually reviewed categories.
  3. Write access is requested separately from the read-only scan.
  4. Before moving a file to recoverable trash, ClearCopies rechecks identity, size and fingerprint against the kept copy.
  5. Each outcome is recorded so interrupted cleanups can resume and results remain auditable.