When metadata disappears, recover the source.
ContentSeal helps ordinary people check a screenshot, repost, crop, compressed image, or direct image URL against lightweight proof receipts, then shows the recovered source, changed-copy risk, and what still cannot be proven.
Exact original
SHA-256 finds unchanged files
Screenshot/repost
pHash recovers source when bytes change
Changed copy
OCR and signals explain risk without fake claims
Product surface
A lightweight recovery workflow for broken provenance.
Seal Source
Issue a lightweight receipt for an original poster, notice, image, or public visual claim.
Open pageRecover Source
Upload a screenshot, repost, crop, compressed image, or paste a direct image URL to find a trusted source.
Open pageTrust Engine
See how integrity, provenance, AI declaration, identity, and lifecycle signals stay separate.
Open pageProof History
Manage receipts, open proof pages, and delete proofs without cluttering the scan workflow.
Open pageDemo Cases
Run the pitch story: original source, screenshot/repost, edited copy, unknown image, and deletion.
Open pageHow the engine thinks
ContentSeal does not call something true. It shows why it can be trusted.
The system separates deterministic proof from weaker context. A file can match a proof receipt exactly, look visually related after a screenshot, contain stripped metadata, include a self-declared issuer, or be expired. Those signals should not collapse into one vague score.
Recovery
SHA-256 catches exact originals; pHash recovers screenshots, crops, and compressed reposts.
Difference
Metadata, OCR, C2PA, classifier, and watermark signals stay explainable.
Lifecycle
Receipts can be active, expired, revoked, replaced, or deleted locally.
Proof page
The proof link survives when platform metadata disappears.
Source Recovered From Screenshot
Visually related to a known proof receipt, but not the original file. Metadata is gone, so users should open the recovered proof page before resharing.
Ready for the demo?