Watermark Remover
Offline Windows tool for removing text, watermarks, and unwanted objects from photos using AI inpainting (LaMa). Brush over what you want gone, keep everything else — including text on clothing.
A local, offline tool for removing watermarks, captions, or unwanted objects from photos. Built around IOPaint running Meta’s LaMa model — the same content-aware inpainting tech behind Photoshop’s “Remove Tool”, but free and running entirely on your own machine.
What it’s for
You took a great photo, but a corner of it has a logo, a date stamp, a watermark, or some background clutter you’d rather not see. Brush over the bits you don’t want, click Run, and it fills them in seamlessly.
How it works
- Run the launcher — a local web app opens in your default browser.
- Drop in a photo.
- Brush over only the parts you want erased. Anything you don’t paint is left untouched, so you can wipe a watermark right next to a logo on a t-shirt without touching the logo.
- Click Run, save the cleaned image.
Once installed, no internet is needed. Nothing leaves the machine.
Install
The repo includes three PowerShell scripts:
| Script | What it does |
|---|---|
setup.ps1 | One-time install — creates a Python venv, installs IOPaint, downloads the LaMa model |
Start-WatermarkRemover.ps1 | Launches the local web UI on 127.0.0.1:8080 |
Install-Shortcut.ps1 | Adds a “Watermark Remover” Desktop shortcut |
Install footprint: ~700 MB (IOPaint + LaMa model weights), one-off download.
Why local?
- Privacy — photos never leave the machine. No upload, no cloud, no telemetry.
- No subscription — the underlying model is open-source.
- Works on a flight — fully offline once installed.
Requirements
- Windows 10 / 11
- Python 3.10+ (one of
Python311,Python312, orPython310) - ~700 MB free disk
- A modern browser