Redact File

Permanently blacks out personal information in PDFs.

Overview

The Redact File node finds personal information on every page and burns black boxes over it. Each page is rasterized before the boxes are drawn, so the redacted values are removed from the file rather than covered by an overlay — they cannot be recovered by selecting, copying, or editing the output.

Detection is done by AI, one page at a time, reading both the page text and an image of the page. Every value the AI reports is located in the page text before it is blacked out, so the boxes land on the real position of the value. When in doubt the AI reports a value, because over-redacting is safer than missing something.

Use it to:

• Share documents that contain personal information
• Strip PII from scanned records before archiving them
• Redact checks, applications, and forms in bulk

Both digital and scanned PDFs are supported. Pages are read from the PDF's own text layer where one is usable, and with OCR where it is not — scanned pages, rotated text, and filled-in form fields. If a page's text layer turns out to be missing something the AI found on the page, such as a signature or handwriting, that page is re-read with OCR automatically.

Parameters

Parameter	Description	Required
File	PDF file to redact (supports expressions)	Yes
Information to Redact	Categories of personal information to black out	Yes
Describe anything else to redact	Additional information to redact, in plain language	No
File Destination	Where to save the redacted PDF	Yes

The redacted file keeps the name of the file it was made from.

Information to Redact

Check each category you want blacked out:

Category	What it covers
Social Security numbers	Full, partial, and masked SSNs
Phone numbers	Telephone and fax numbers
Email addresses	Email addresses
Names	Names of people, including signatures and initials — company and institution names are not covered
Addresses	Street addresses, along with the city, state, and postal code that form part of them
Dates of birth	Dates of birth only, not other dates
Credit card numbers	Card numbers, expiration dates, and security codes
Bank account & MICR (checks)	Account and routing numbers, including the full MICR line on a check
Government IDs (license, passport, EIN)	Driver's license, passport, visa, and taxpayer ID numbers
Other personal information	Any other identifier tied to a person: policy, member, claim, and medical record numbers, usernames, IPs, plates

At least one category must be checked, or a description filled in below.

Describe anything else to redact

Anything the categories do not cover, described in plain language:

Redact the policy number and the claim reference in the header

Settings

Setting	Description
Execution Mode	Once per item (default) or Once
Output Mode	How to output results when running once
Batch Size	Items to process concurrently (default 5)
Stop on Error	Stop workflow on failure

Output

{
  "file": {
    "type": "fileData",
    "name": "statement.pdf",
    "fileInfo": { "type": "..." }
  }
}

Access in expressions:

• File object: {{$item.data.file}}

Examples

Redact Before Sharing

[Google Drive Trigger] → [Redact File] → [Copy File (shared folder)]

Redact Emailed Attachments

[Lido Mailbox Trigger] → [Split (attachments)] → [Redact File] → [Copy File (archive)]

Tips

• Billing is per page, so a 10 page document costs 10 actions
• Documents are limited to 500 pages
• The output pages are images, so the redacted PDF has no text layer and is not searchable
• If any page fails — for example, a value the AI reported cannot be located on the page — the whole file fails and nothing is uploaded, so a partially redacted document is never produced
• This is a long-running operation — processing time depends on page count
• Zoom in on the output to confirm coverage: the boxes are part of the page image, not an overlay
• Connect error output to handle redaction failures