Open an image. Draw zones. Redact. Export. Nothing leaves your machine.
RedactIT is a lightweight PyQt6 GUI tool for redacting sensitive information from images — IPs, credentials, names, locations, anything you want gone before sharing a screenshot. It runs entirely offline. No uploads, no telemetry, no cloud.
Built for Linux with a KDE-native dark theme (KlassyDark), it supports OCR-assisted text detection, multiple redaction styles, and full per-zone customisation — all in a single Python file.
| Method | Description |
|---|---|
| Blackout | Solid fill — customisable colour |
| Whiteout | Solid white fill |
| Blur | Gaussian blur — variable radius |
| Pixelate | Mosaic/pixelation effect |
| Distort | Sine-wave spatial warp |
| Clone Stamp | Replaces region with content from a chosen source point |
Every redaction zone is independently configurable:
- Method — switch any zone to any method at any time
- Intensity — controls blur radius, pixelation block size, distort strength
- Padding X / Y — expand the zone outward from the drawn region
- Opacity — blend the redaction against the original (0–100%)
- Corner Radius — rounded rectangle clipping mask
- Border Width — optional visible outline (0 = invisible)
- Tolerance — feathered/soft edge fade on the redaction boundary
- Fill Colour — picker for Blackout/Whiteout fills
Run the built-in Tesseract OCR scan and every detected word appears as a clickable chip in a strip below the canvas. Click any word chip to instantly create a redaction zone around it. OCR highlights on the canvas are only visible in Select mode — the canvas stays clean otherwise.
- Draw mode — drag to place a zone anywhere on the image
- Select mode — click a zone to load it into the editor; click empty space to deselect
- Clone Source mode — click to pick a source region for Clone Stamp zones
- Scroll-wheel zoom (10–500%), synced zoom slider in the bottom bar, Fit button
- Non-destructive — original file is never modified; preview updates live as you edit
- Undo / Redo — 60-level stack
- Alpha-transparent images — PNG, WEBP, TIFF with transparency are opened, edited, and exported with their alpha channel fully intact. JPEG exports flatten against white (JPEG spec limitation)
- Export log written to
~/.log/redactit.log
Python 3.10+
tesseract-ocr
PyQt6 >= 6.4.0
Pillow >= 10.0.0
numpy >= 1.24.0
opencv-python >= 4.8.0
pytesseract >= 0.3.10
Installs pip dependencies, the script, a shell launcher, a .desktop entry for your application menu, and an icon:
git clone https://github.com/DitherZ/RedactIT.git
cd RedactIT
python3 setup.pyThen launch from your application menu or terminal:
redactitgit clone https://github.com/DitherZ/RedactIT.git
cd RedactIT
python3 setup.py deps-only
python3 redactit.py# System dependency
sudo apt install tesseract-ocr
# Python dependencies
pip install --break-system-packages -r requirements.txt
# Run
python3 redactit.pypython3 setup.py uninstall1. Open — File → Open Image, or click 📂 Open in the toolbar
2. Scan OCR — Click 🔍 Scan OCR to detect text; word chips appear below the canvas
3. Draw zones — Drag on the image in Draw mode to place a redaction zone
— OR click a word chip to auto-zone a detected word
4. Edit zone — Select a zone in the left panel; edit it in the Zone Editor on the right
5. Preview — Preview updates automatically (400ms debounce); click Refresh Preview to force
6. Export — Click 💾 Export or File → Export to save the redacted image
| Shortcut | Action |
|---|---|
Ctrl+O |
Open image |
Ctrl+S |
Export redacted image |
Ctrl+Z |
Undo |
Ctrl+Shift+Z |
Redo |
Ctrl+Q |
Quit |
Scroll wheel |
Zoom in / out on canvas |
RedactIT uses Tesseract via pytesseract to scan the image locally. No image data is sent anywhere.
After scanning:
- Each detected word with confidence ≥ 30% appears as a chip in the strip below the canvas
- Amber dotted highlights appear on the image in Select mode
- Clicking a chip (or a highlight) creates a redaction zone around that word using the current method
- The OCR scan result is reusable — you can keep clicking chips to zone multiple words
- Fully offline — OCR and all processing run locally via Tesseract and Pillow/OpenCV
- Non-destructive — source image is never modified; a copy is processed on export
- No telemetry — no network calls of any kind
- Export log —
~/.log/redactit.logrecords export paths and zone counts locally
RedactIT/
├── redactit.py # Main application (single-file)
├── setup.py # Installer: deps, launcher, .desktop, icon
├── requirements.txt # Python package requirements
├── screenshot.png
└── README.md
| Version | Summary |
|---|---|
| 2.1.0 | KlassyDark colour scheme; zone border width control; per-zone ✕ in sidebar |
| 2.0.0 | Full UX overhaul; fixed C++ deleted-object crash; OCR chip strip; zoom slider; draw-mode-default; non-destructive alpha-safe preview |
| 1.0.1 | Fixed PyQt6 QAction overload error in menu builder |
| 1.0.0 | Initial release |
Issues and PRs welcome. This is a personal-use tool so scope is intentionally narrow — redaction, not annotation. If you're adding a new redaction method, the RedactionEngine._process() method is the right place; add the method name to REDACT_METHODS and a colour to METHOD_COLORS.
MIT — see LICENSE.