---
license: apache-2.0
language:
- ar
size_categories:
- 1M<n<10M
---
# AraBooks

A dataset comprising OCR-scanned Arabic language content from a vast collection of books, intended to bolster development and research in the Arabic languages. It includes texts from a variety of sources, some translated into Arabic while others originally authored in the language.

This dataset may contain noise due to the low quality of PDF files.

## Stats behind dataset:
  - total number of tokens (all books combined): ~ **555,643 tokens**.

**Disclaimer:**

The dataset provided herein consists of scanned books and is intended strictly for research and educational purposes. The content of these scanned books may be subject to copyright protection and other intellectual property rights.

Users of this dataset bear sole responsibility for ensuring compliance with applicable copyright laws and regulations. The creator of this dataset offers no assurances regarding the accuracy, completeness, or suitability of the data for any specific purpose.

Moreover, the creator of this dataset disclaims any liability for loss, damage, or legal repercussions resulting from the use or misuse of the provided data.

By accessing and utilizing this dataset, you agree to indemnify and absolve the creators of the dataset from any claims, damages, or liabilities arising from your use of the data.