Persian Log2Vis / Arabic Log2Vis / A PHP project to convert logical UTF8 Persian and Arabic characters to visual.

A lightweight Rust library for removing Arabic diacritics

Text-to-image conversion (OCR) for Pashto and Chinese, with a view towards comprehensive, multi-lingual OCR

RTL fixes for Arabic-script languages in Codex through ChatGPT Desktop on Windows

Pashto Natural Language Processing Toolkit

PALI: Language identification for Perso-Arabic Scripts

Beautifies Arabic, Persian, Urdu, Pashto and other right-to-left (RTL) languages

ایجاد متن های ساختگی دری و پشتو

A LoRA-fine-tuned Whisper system that turns Pashto speech into text, deployed as a real-time Streamlit web app.

Hadith collection in Pashto language, developed by flutter.

Check how many Pashto words you can type in one minute, check you typing speed in Pashto language, imporve your Pashto typing speed, learn how to type in Pashto.

This is an extension to set dir="auto" in content wrappers to make them right-aligned when the language is an RTL language (like Persian, Arabic, ...)

Convert numbers to legal Afghan Afghani (AFN) wording in Dari and Pashto. Perfect for invoices, banking, and official documents.

Website for reading Quran with Pashto translation

X Series 2 Webfont for Arabic Script

This repo contains Pashto fonts hosted by pashtology.com

How well do frontier AI models actually speak Pashto? A native-authored capability and safety benchmark for a language of 40+ million people.

پښتو باران يوه پښتو ليکبڼه ده چې په ځانګړې توګه د پښتو ژبې وېبپاڼو لپاره د نازنين او اېکس بي کيهان ليکبڼو تر اغېز لاندې ډيزاين شوې ده تاسو کولی شئ ياده ليکبڼه هرځای کې له وړيا سوداګريزې کارونې جواز سره د پښتو، اردو، عربي، فارسي، کهوار، سرائيکي، ازربايجاني تورکي، عثماني تورکي، له کي، مازروني، پنجابي......ژبو لپاره وکاروئ.

An open-source multimodal corpus for Pashto, integrating text, images, audio, video, metadata, and knowledge graphs to advance AI, NLP, vision-language research, and digital preservation.

Fine-tuned Whisper Large-v3 + VITS2 for Pashto & Dari — LoRA, CTranslate2 INT8, WER 12.29%, trained on 1.5M+ audio samples, deployed on Modal.