Safety & privacy
Updated August 12, 2026 · min read
File Metadata vs File Contents: What a OneDrive Scanner Can See
Distinguish filenames, paths, sizes, hashes, and sharing state from the bytes inside a document or media file.
Saying that a scanner does not download contents is meaningful, but it does not make the remaining data anonymous. A path can reveal a client name, a filename can reveal a diagnosis, and sharing fields can reveal relationships.
Create a data map before connecting: fields requested, purpose, storage location, retention, access, export, and deletion. Limit the scan record to what exact grouping and review genuinely require.
A practical checklist
- 1List every file field the scanner requests.
- 2Separate cloud API metadata from file-body transfer.
- 3Check storage duration and who can access scan records.
- 4Verify that exact groups use size plus supported fingerprint.
- 5Delete retained scan data when it is no longer needed.
Understand the situation
Saying that a scanner does not download contents is meaningful, but it does not make the remaining data anonymous. A path can reveal a client name, a filename can reveal a diagnosis, and sharing fields can reveal relationships.
A permission label describes technical access, not the provider’s entire security posture. Review the exact scope, token handling, retention, operator access, and purpose of every data field before connecting an account.
Use a controlled workflow
Create a data map before connecting: fields requested, purpose, storage location, retention, access, export, and deletion. Limit the scan record to what exact grouping and review genuinely require.
Use least privilege in stages: read-only access for inventory and review, followed by separate write consent only if the user chooses an approved cleanup. A report-only workflow should remain possible.
- Separate cloud API metadata from file-body transfer.
- Check storage duration and who can access scan records.
- Verify that exact groups use size plus supported fingerprint.
Make the final decision
Choose a tool whose product promise matches the data flow. Exact matching can use size and OneDrive’s fingerprint; previewing or semantic analysis would require a different, more invasive architecture and disclosure.
Metadata is not harmless just because it excludes file bodies. Names, paths, sizes, account IDs, and sharing state can reveal sensitive business or personal context and need deliberate protection.
Limits and risks to check
- — Metadata can be commercially or personally sensitive.
- — Thumbnail or preview features may process more than basic metadata.
- — Logs and exports can create extra copies of metadata.
- — A vague “no download” claim may omit server-side fetch behavior.
Official references
Frequently asked questions
Is a hash file content?
It is a content-derived metadata value, not the file body. It still needs protection and clear purpose.
Can metadata reveal document text?
Basic item metadata does not contain the document body, but filenames, paths, previews, or extracted properties can reveal context.
Does ClearCopies inspect photo pixels?
No. Its MVP exact scan uses file metadata and does not perform visual-content analysis.
Scan first. Decide with evidence.
ClearCopies reads supported cloud-drive metadata and groups exact copies by byte size plus a provider-supplied content fingerprint. Original file bodies are not downloaded for the scan. You review the result and export a plan before any separate write step.