Install Tesseract OCR

niuniu’s Reading Acceleration uses Tesseract OCR locally to turn text inside screenshots, scans, and images into searchable text. It all runs on your own machine — images are never uploaded.

Tesseract is a standalone system program and is not bundled with niuniu. If it isn’t installed, niuniu shows an “OCR engine not found” notice and sends you here. Follow the section for your operating system below — it usually takes 1–2 minutes.

You only need to install it once. After that niuniu auto-detects the tesseract command; no reboot required.

1. Install by platform

macOS

The easiest path is Homebrew:

brew install tesseract

This installs the Tesseract engine with English (eng) and orientation-detection (osd) data included. To recognize Chinese or other languages, add the full language pack:

brew install tesseract-lang

No Homebrew yet? Run the official one-line installer first — /bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)" — then come back and run the commands above.

Linux

Debian / Ubuntu family:

sudo apt update
sudo apt install -y tesseract-ocr
# Language packs: Simplified + Traditional Chinese (English eng ships with the core package)
sudo apt install -y tesseract-ocr-chi-sim tesseract-ocr-chi-tra

Fedora / RHEL / Alibaba Linux family:

sudo dnf install -y tesseract
sudo dnf install -y tesseract-langpack-chi_sim tesseract-langpack-chi_tra

Arch / Manjaro:

sudo pacman -S tesseract tesseract-data-eng tesseract-data-chi_sim tesseract-data-chi_tra

Kylin / UOS and similar domestic distros are usually Debian- or RHEL-derived — use the matching command group; the package names are the same.

Windows

Windows has no official installer; the community-maintained UB Mannheim build is the de-facto standard:

1. Open the UB Mannheim Tesseract download page and download the latest tesseract-ocr-w64-setup-*.exe (64-bit).
2. Run the installer. At the “Additional language data” step, tick the languages you need — at minimum Chinese (Simplified) and Chinese (Traditional); English is included by default.
3. Keep the default install path C:\Program Files\Tesseract-OCR.
4. Important: on the “Select Additional Tasks” step, enable “Add to PATH”. If the installer doesn’t offer it, manually add C:\Program Files\Tesseract-OCR to your system Path afterward.

When it finishes, open a new PowerShell window (so the updated PATH takes effect) before running the verification below.

2. Language packs

Tesseract’s recognition is driven by language data files (*.traineddata), one per language:

Language	Data code
English	eng
Simplified Chinese	chi_sim
Traditional Chinese	chi_tra
Orientation / script detection	osd

• If you installed only the engine without a Chinese pack, Chinese recognition will fail or come out garbled — be sure to add chi_sim (and chi_tra if needed) using the command for your OS above.
• Language data lives in Tesseract’s tessdata directory by default. If you put the data elsewhere, point the TESSDATA_PREFIX environment variable at that directory.

3. Verify the install

Open a terminal (on Windows, the freshly opened PowerShell) and confirm it works:

tesseract --version

A version line like tesseract v5.x.x means the engine is ready. Now check the language packs:

tesseract --list-langs

The output should include eng, chi_sim (and chi_tra if you installed it). When all of them appear, you’re set.

Version requirement: Tesseract 4.0 or newer is recommended (5.x is best). Older Linux distro repos may ship 3.x, whose accuracy is much worse — prefer the commands above to get a current build.

4. Back in niuniu

Once the dependency is installed:

1. Return to niuniu and trigger Reading Acceleration / OCR again (or click “Re-check” wherever the “OCR engine not found” notice appeared).
2. niuniu re-detects the tesseract command on your system; once found, recognition works.
3. If it still says not installed, the cause is almost always PATH not taking effect — fully quit niuniu and your terminal, reopen them, and try again.

5. Troubleshooting

Installed it, but niuniu still says it’s missing? Almost always tesseract isn’t on your PATH. Run tesseract --version in a terminal: if the terminal also reports “command not found,” PATH isn’t set up (on Windows, double-check the “Add to PATH” step). Reinstall or add the PATH manually per your OS section.

Chinese comes out blank or garbled? The Chinese language pack is missing. Run tesseract --list-langs to check for chi_sim; if it’s absent, add the language pack from your OS section.

--list-langs finds no languages at all? The language-data directory is wrong. Point the TESSDATA_PREFIX environment variable at the tessdata directory that actually holds your *.traineddata files.

Next steps

• Install Personal desktop — installing niuniu itself and its dependencies
• FAQ — other dependencies and troubleshooting