OCR Screen Translator

Capture any part of your screen, extract text with OCR and translate it instantly.

---

About project

This is a lightweight Windows desktop tool built for one fast workflow:

1. Select an area on screen.
2. Run OCR on the captured image.
3. Translate the extracted text.

No tab switching, no copy-paste between multiple apps.

preview:

intro.mp4

Features

• Global hotkey capture from anywhere in Windows
• In-memory OCR pipeline (no temporary screenshot files required)
• Instant translation using Google Translate (default)
• Translator architecture ready for additional providers
• Copy and save actions for fast output handling
• Desktop executable build with PyInstaller

Keyboard shortcuts

Shortcut	Action
Ctrl+Alt+T	Capture area and translate
Ctrl+Alt+C	Copy output
Ctrl+Alt+S	Save output to file

How it works

1. The app starts and requests Administrator permission on Windows.
2. Global hotkeys are registered using the keyboard library.
3. You capture a region with the snipping overlay.
4. OCR extracts text using Tesseract.
5. The selected translator returns translated output to the UI.

Tech stack

Layer	Technology
Desktop bridge	Eel
OCR	Tesseract + pytesseract
Screen capture	mss + tkinter
Translation	requests + Google endpoint
Hotkeys	keyboard
Packaging	PyInstaller

Quick start

1) Prerequisites

• Windows
• Python 3.10+
• Tesseract OCR installed
  • Recommended installer: https://github.com/UB-Mannheim/tesseract/wiki

2) Install dependencies

pip install -r requirements.txt

3) Run app

python main.py

Note: On startup, Windows may show a UAC prompt so global hotkeys can work system-wide.

Build executable

python build.py

Trusted GitHub Releases workflow

To improve user trust for downloaded binaries:

1. Build with safer defaults:

python build.py

2. Upload these generated metadata files from dist/ with your release assets:

• main.zip (or your custom zip name)
• RELEASE-SHA256.txt
• RELEASE-NOTES-TEMPLATE.md
• SIGNING-COMMAND-TEMPLATE.txt

3. In release notes, include checksum verification steps (PowerShell):

Get-FileHash .\ScreenOCRTranslator.exe -Algorithm SHA256

4. Prefer code-signed binaries for fewer SmartScreen warnings. SIGNING-COMMAND-TEMPLATE.txt contains ready-to-edit signtool examples.

Cross-platform standalone (macOS)

Yes, it is possible, but build must run on macOS.

• Windows build produces Windows binaries only.
• For macOS users, build on a Mac machine (or macOS CI runner) with the same source code.
• Typical macOS output is .app (and optionally .dmg/.zip).
• To avoid Gatekeeper warnings on macOS, use Apple Developer code signing and notarization.

Suggested release assets:

• ScreenOCRTranslator-win-x64.zip (built on Windows)
• ScreenOCRTranslator-macos-arm64.zip (built on Apple Silicon Mac)
• ScreenOCRTranslator-macos-x64.zip (built on Intel Mac)

GitHub Actions release automation

This repository includes a workflow at .github/workflows/release-build.yml that:

• Builds standalone ZIP assets for Windows and macOS
• Generates per-platform SHA256 checksum files
• Uploads assets automatically when a Release is published

Extend with another translator

1. Open screen_translator/translator.py.
2. Create a class that extends BaseTranslator.
3. Implement translate.
4. Register the class in TRANSLATORS.

The UI list is populated dynamically.

License

This project is licensed under the MIT License.