Topkill/tianruoocr

TianRuo OCR v6 – a lightweight Windows OCR + translation tool

What it is

• A Windows desktop application that lets you capture text from the screen, recognise it (OCR) and translate it to another language. It builds on the earlier tianruoocr project and adds bug‑fixes and a few new features.

Key capabilities

Area	How you use it	What you get
Text capture & OCR	• Double‑click the tray icon or press F4 for a normal scan. • Use the silent mode (tray menu or shortcut) to copy the recognised text directly to the clipboard without a pop‑up.	Recognises printed text, tables and (if the chosen backend supports it) handwritten text.
Translation	• After OCR, click the toolbar button to translate. • Right‑click the translation window to change source/target languages and re‑translate. • Quick translate (default F9) works on any selected text: with fast mode the result is copied to the clipboard (and pasted if possible); with window mode a small result window appears. • Screenshot‑translate: trigger from the tray or a shortcut, the captured image is OCR‑ed and the translation shown (or silently copied). • Clipboard‑listen mode automatically translates anything you copy.	Supports multiple online translation services (Google, Bing, Microsoft, Yandex, etc.) and AI‑based translation endpoints that accept prompts.
Table handling	Choose Baidu or Tencent table‑OCR services; after a successful scan you can export directly to Excel or paste with Ctrl + V.	The extracted table retains its data; formatting may differ between the two export methods.
Handwriting	Select Baidu Handwriting, Tencent High‑Precision, or an offline model that supports handwriting.	Hand‑written characters are turned into editable text.

Underlying engines

• Online OCR services – Baidu, Tencent, etc. (require API keys you obtain yourself). Some services are free up to a quota; the app warns you to monitor usage.
• Offline OCR back‑ends – You can pick one of three bundled options: PaddleOCR, PaddleOCR2, or RapidOCR. These run locally and need no internet connection.
• Translation back‑ends – All are online; you configure API keys in the Settings → Translation Interface page. AI‑style translation (prompt‑based) is also supported via the same “AI interface guide”.

Configuration highlights

• All shortcuts are configurable; defaults are F4 (OCR) and F9 (translate).
• In Settings → General you can toggle silent OCR, fast translation, and screenshot translation behaviours.
• The “Auto‑translate after text change” option lets you set a debounce time (milliseconds) and optionally whitelist/blacklist specific translation services using a simple comma‑separated syntax (e.g., 1000,Bing,Bing2).
• API keys are entered under the Key sections; the app does not ship any keys.

Known limitations

• The process name shown in Task Manager changes to the elapsed time after a recognition task, which can be confusing.
• Some emoji characters are not rendered correctly in the UI and appear as squares.
• Tested on Windows 10; other Windows versions may exhibit quirks.

Who might find this useful?

• Users who need quick on‑screen text extraction (e.g., reading PDFs, games, or chat windows) without opening a full‑blown OCR suite.
• People who frequently translate snippets of text and want a single hot‑key workflow.
• Developers looking for a ready‑made Windows front‑end that can be pointed at various OCR/translation APIs.

Getting started

1. Download the latest release (or build from source).
   
2. Run the program; it will appear as a tray icon.
3. Open Settings → Key to add any required API keys for the OCR or translation services you plan to use.
4. Use F4 to capture text, F9 to translate, or explore the tray menu for silent and screenshot modes.

Acknowledgements The project reuses code from the original tianruoocr repo and incorporates fixes and additional interfaces from several other open‑source tools (e.g., STranslate, GTranslate, WeChatOcr, pot‑desktop).