Undocumented hidden PDF text layers silently break redactions and page-citation accuracy

Summary

Litigation teams downloading PDFs from eDiscovery platforms (e.g., NextPoint) are discovering that the produced documents carry an unconventional, undocumented text layer — text placed in invisible 1pt white font at the edge of page 1 rather than spatially aligned OCR. Two severe consequences follow: (1) search results report the wrong page location, so a lawyer preparing a hearing who relies on 'page 47' in a citation is pointed to the wrong page, and source-anchoring of quotes collapses; (2) far more dangerously, a paralegal performing manual redactions by highlighting visible text leaves the hidden layer intact, so confidential or privileged text survives into the 'redacted' production. The platform vendor confirms this is intended behavior and it appears in no public documentation. Today there is no tool that inspects a PDF for stray/hidden text layers, warns the reviewer, or verifies that a redaction actually removed every occurrence of the string. Teams currently discover privilege leaks only after production — the worst possible time. This is a high-stakes, genuinely unsolved integrity problem that directly undermines any product claiming to read case files and anchor citations to the source.