RedactSense

RedactSense is an offline desktop app that uses on-device AI to detect and permanently remove sensitive data from PDFs, Word files and images, including metadata, without uploading anything to a cloud server.

RedactSense is a desktop application that finds and permanently removes sensitive information from PDFs, Word documents and images. Its defining trait is that everything happens on the user's own machine: the AI models run on the local CPU or GPU, no files are uploaded, and the app works without an internet connection. That design targets professions where sending client or patient material to a cloud service is ruled out by policy, such as law, healthcare, HR and accounting. The workflow has four stages. A user opens a file, the on-device models scan it for confidential items (the homepage demo shows names, Social Security numbers and card numbers being flagged), and the results are presented for review. The reviewer can add anything the models missed or restore items that should stay visible. On export, the data is removed from the file itself and from its metadata, rather than hidden under a black rectangle that leaves the original text searchable. That distinction is the central argument of the product, since cosmetic masking in a PDF is a known source of accidental disclosure. Detection is described as context-aware instead of pattern-only, so it is meant to catch natural phrasing and not just formats like email addresses. The vendor says tables, scanned pages and multi-column layouts are handled locally through vision and text models. It also publishes figures for detection accuracy and speed against manual work, though these are the company's own claims and no methodology is given on the homepage. HIPAA and GDPR readiness are listed as badges. Licensing runs from a free tier that exports only the first two pages of each document, through per-device Individual and Team plans, to custom Enterprise volume deals with admin device management and an SLA. Compared with cloud redaction services, it trades server-side convenience and collaboration for stronger data-residency guarantees. Compared with manual redaction in a PDF editor, it automates the detection step while keeping a human review stage. It suits privacy-bound professionals and compliance teams who need repeatable redaction without a vendor ever touching the documents.