Tesseract OCR CLI
Tesseract OCR is a CLI for multilingual scans: select language datasets and compare recognition on a sample page.
- 77,377 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- The CLI reads images and writes recognized text or documents to disk.
Install
Manual install
Open https://tesseract-ocr.github.io/tessdoc/Installation.html and select the OS package. Show required language data. Recognize one image copy using the README syntax, save to a new basename and compare with the source.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The CLI reads images and writes recognized text or documents to disk.
This is third-party code. Review the repository files before installing.
What it does
Tesseract OCR is a CLI for extracting text from PNG, JPEG and TIFF images. Users supply a file, select a language and receive recognized text or a document with a text layer. The project includes the `tesseract` command and a library, without a graphical application. Pages can use several language datasets together. Page segmentation controls how an image is divided into text regions. TSV, hOCR, ALTO and PAGE outputs retain structure for subsequent processing by other tools.
Who it is for. Document-archive specialists recognizing page images for search or data processing.
Good fit when
- An archive contains TIFF scans and an operator creates text versions.
- An analyst needs TSV recognition data for further parsing.
- An editor creates a text-layer PDF from a page image.
- A page mixes languages and the operator selects several traineddata sets.
Not a fit when
- A user wants to pass PDF directly as input to this engine.
- Staff need a complete desktop interface for reviewing and editing pages.
- The process needs table or chart semantics without further result analysis.
Example request
Recognize a scan copy with Tesseract, save text separately and show a passage beside the source image.Limitations
Each chosen language requires matching `traineddata` files; legacy-engine mode needs compatible data from tessdata. Leptonica does not open input PDFs, so page images must be prepared before recognition. Results depend on image quality, and the author recommends improving source pages before OCR. No graphical interface is bundled; reviewing and correcting recognized text requires a separate editor or application built around the engine.
How to disable. Stop `tesseract` and remove it from the pipeline. Output documents remain; decide separately whether to retain language data when removing the program.
FAQ
How do I install the program and languages?
Choose your operating-system package at https://tesseract-ocr.github.io/tessdoc/Installation.html and install data for the required languages. Start with one page after setup; compiling from source is an alternative.
How do I recognize one image?
The guide gives `tesseract imagename outputbase`. Replace `imagename` with the image path and choose a new output basename. English and text output are defaults.
How do I choose a language?
`-l LANG` accepts a three-letter code. `tesseract images/eurotext.png - -l eng` prints English text to the terminal; install the relevant traineddata first.
Can I combine two languages?
Yes. `tesseract images/eurotext.png - -l eng+deu` combines English and German. Language order can affect output, so compare it on a sample page.
Where does language data go?
The guide identifies the `tessdata` directory and `TESSDATA_PREFIX`. Match your installed traineddata location to the configured path before retrying.
Discussion
No questions or reviews yet.
Related
CLI toolsDocling is a CLI for mixed document archives, bringing PDFs, office files and spreadsheets into a shared format for AI search.
MinerU is a CLI with a document library: search and revisit an already parsed report using its local cache.
PaddleOCR agent skills
PaddleOCR Agent Skills
Two official PaddleOCR skills: text recognition from images and PDFs, and parsing complex documents into Markdown
Document extraction engine for PDFs, Office files and scans, shipped as an agent plugin, MCP server and CLI
A local CLI extracts text from documents and scans, searches archives and prepares source material for an agent’s answer.
MCP PDF for large documents: the agent searches by meaning and keyword across one file or a folder and reads only the pages, tables and scans it needs
README in short
The README covers neural and legacy engines, language data and image formats. Installation uses packages or a source build. C/C++ APIs support application integration.