Before anything is summarized or compared, each document is read (text-recognition / OCR for scans), personal and protected information is removed, and the result is stored as structured data in a central repository that every later stage draws from. A residual scan and a person confirm nothing identifying remains.
Documents processed
9
text-recognized (OCR)
Items removed
214
100% of those detected
Residual scan
0
identifiers remaining
Checkpoint
Pending
mediator confirmation
How redacted data is used downstream. Each document is turned into structured fields (e.g., "retirement account = $180,000, source: 401(k) statement") held in a central repository. Later stages read those fields directly — the figures feed Summarize, Gap Analysis & Alignment, and Resolution with full provenance, rather than the documents being mere attachments.
What was removed, by category
Category
Action
Items
Names of the parties
→ neutral labels Party A / Party B (consistent)
86
Children's names
→ Child 1 / Child 2 (ages kept)
22
Government identifiers (Social Security, licence)
Removed
7
Account numbers
Masked (last four kept)
31
Addresses
Removed (county kept)
18
Dates of birth
→ age only
12
Employers & contact details
Removed
38
Total
214
Documents processed — titles & descriptions only; no document content is shown
Document
Description
Topic
Removed
Status
Statement of Net Worth — Party A
Sworn assets, income, expenses, liabilities
Finances · Support
48
Redacted
Statement of Net Worth — Party B
Sworn assets, income, expenses, liabilities
Finances · Support
51
Redacted
Retirement & bank statements
Account balances, last four kept
Finances
39
Redacted
Pay stubs & tax return
Wage and income figures
Support
44
Redacted
Parenting calendar & notes
Schedule; children → Child 1 / Child 2
Children
32
Redacted
Residual-identifier scan & checkpoint
Passed — 0 residual identifiers. A second pass searched the redacted text and structured fields for names, numbers and patterns that could re-identify a person; none found.
Human checkpoint before proceeding
The mediator reviews the summary and residual scan and confirms before the matter advances to Summarize. Nothing un-redacted is ever stored or shared.
Text-recognition + redaction run June 16, 2026 · 214 items · residual scan passed (0) · structured data written to the matter repository · awaiting mediator sign-off · job hash and actor logged.
Reports available
Summary Report
A one-glance dashboard of this step
Detailed Report
Full inputs → process → references → outputs, with the audit chain
De-identification only. Redaction changes how information is labeled, never the substance of the figures used in the comparison.
JusticeX is a software platform — not a law firm or mediator. It provides objective, mathematical comparison (and optional public legal precedent) for reference only — no legal advice, interpretation, or argument guidance; the two parties reach their own agreed view. Full legal & UPL disclaimer. Synthetic demo — no real person, figure, or document.
JusticeXObjective comparison