Data pillar · verified-clean document redaction

Redact anything in minutes, not hours.

Upload a document. ShroudCloud finds every name, number, and secret. You review each one, then export a verifiably clean file with a complete audit trail.

Manual redaction is slow, and a single missed name — or a black box that lifts right off the text underneath — can become a headline. The tools you already have either miss too much or make you hunt through the document line by line.

The workflow

1. Upload

PDF or Word. Files are processed and destroyed at 24 hours by lifecycle rule — nothing persists.

2. Review

Every proposed redaction with page previews. Accept, reject, or draw manual boxes. You stay in control.

3. Export verified

A verifiably clean file, plus a JSON audit record of every entity, action, and timestamp rendered to a PDF report.

The verification pass that never gets cut

A redaction product that leaks once is finished. That's why every export is re-run through the detector: if any detected entity survives, the export fails closed. Copy-paste resistant, metadata scrubbed, XMP stripped.

Detect

Presidio + custom recognizers for the entities that matter: JWTs, AWS keys, bearer tokens, private key blocks.

Redact

True redaction, not a black box. Text removed at the content layer, coordinates applied, metadata scrubbed.

Verify

Detection re-run on every output. Fail closed on any survivor. This is the headline feature, never cut.

FAQ

What file types are supported?

PDF and DOCX at launch, with metadata scrubbing on both. OCR for scanned PDFs is on the roadmap but explicitly not in the MVP.

What does the audit trail include?

A JSON record of every entity with its action (auto, accepted, rejected, manual), the actor, and a timestamp, rendered to a PDF report with the export.

What about pricing?

No pricing announced yet. Early access users lock in founding terms.