PaintTraceWipe

Who it's for

Anonymity-seeking Windows users, journalists, whistleblowers and OSS maintainers who quick-edit screenshots in MS Paint/Photos; today they rely on ExifTool, hex editors or nothing, all of which miss the steganographic GUID

The problem

legal — the invisible GUID can fingerprint and deanonymize a leaked screenshot or forum post, creating doxxing or subpoena exposure

How to build it

Local Tauri/WASM drag-drop app (Windows-first) with optional CLI; batch folder support, offline only

How it makes money

Privacy pros and freelancers pay $11 one-time or $5/mo for unlimited batch+CLI; free tier caps at 3 images/day and leaves a 2-pixel corner mark so clean output requires payment and free EXIF tools cannot substitute

Why it doesn't exist yet

The watermark scheme was reverse-engineered only in the last 24 h and is still undocumented by Microsoft; incumbents chase EXIF/browser trackers or AI SynthID marks and have no incentive to strip a rival OS fingerprint, leaving a narrow indie window

First users

Commenters and upvoters on the fresh 495-pt HN reversing thread who just realized their recent Paint exports are fingerprinted and want an immediate cleaner

Build size

1 person x 5 weeks: detector+stripper for the published Paint/Photos GUID stego patterns, simple GUI+CLI, offline verification suite; excludes non-MS watermarks and mobile OSes

Biggest risk

Microsoft ships a Windows update that alters or drops the watermark format before the tool gains traction, or a zero-cost FOSS clone captures the entire HN audience first

Conditions for a hit (all 3 required)

• Accepts PNG/JPEG, detects the specific MS Paint/Photos GUID steganographic pattern and extracts the GUID (if present) within 10 seconds
• Outputs a cleaned image with the watermark payload overwritten by zeros or noise, plus a one-page PDF/JSON report of the original GUID and bit offsets
• Runs fully offline after initial load (verifiable by airplane-mode or packet capture showing zero network calls)

How it's judged (in 6 months)

GitHub ≥750 stars or Product Hunt daily top 5 or ≥2 000 paid license sales(judgment date 2027-02-25)

AI self-confidence 47/100 — self-reported likelihood of meeting the criterion, not a business success rate