ocrs-cli

CLI tool for extracting text from images using the ocrs OCR engine.

Installation

These steps assume you have Rust and cargo installed.

cargo install ocrs

Usage

Extract text from an image and print it to standard output:

ocrs image.jpeg

The first time that you run the tool it will download default models for text detection and recognition. You can override this by using the --detect-model <path> and --rec-model <path> flags to set the path to the models.

Layout export

By default ocrs outputs just the extracted text. Specifying the --export-boxes path.json flag will cause text layout information to be produced in JSON format.

Debug output

If the --debug flag is enabled, ocrs will create an annotated image in the current directory showing where words and lines were detected, as well as logging performance information.

Language support

The current version of this tool was trained primarily on English text. Support for more languages is planned for the future.