ali2000hos/persian-writing
Professional Persian (Farsi) writing & RTL documents skill for any AI — نگارش فارسی حرفهای برای هوش مصنوعی
What it solves
It addresses two primary failures of AI-generated Persian (Farsi) text: prose that sounds artificial ("AI-tell" patterns) and technical document formatting issues where Right-to-Left (RTL) layouts, fonts, and orthography (like ZWNJ and Persian characters) are broken or incorrectly aligned.
How it works
The project provides a comprehensive "skill" package that can be integrated into AI agents (like Claude, Cursor, or ChatGPT) or used as a standalone Python CLI. It employs a multi-layered approach:
- Register Detection: A 6-step process to classify the required tone (formal, administrative, colloquial, or scientific).
- Humanization: Identifies and fixes 18 common Persian AI-writing patterns to make prose sound more natural.
- Orthography & Cleanup: Uses a dictionary-backed toolkit to fix spell-checking, ZWNJ (half-spaces), and character corrections (Persian vs. Arabic characters).
- RTL Document Engineering: Provides specific recipes and verification scripts for Word, PDF, PowerPoint, HTML, and Excel to ensure correct RTL alignment and font rendering.
- Academic & SEO Modules: Includes specialized rules for scientific citations, SEO copywriting, and social media formatting.
Who it’s for
- AI users and developers who need professional, human-sounding Persian content.
- Content creators and SEO specialists working in the Persian language.
- Academic researchers needing faithful Persian paraphrasing and formatting.
- Developers building AI agents that must output high-quality RTL documents.
Highlights
- Universal Compatibility: Works as a Claude skill, a plugin for Claude Code, a knowledge file for GPTs/Gemini, or a standalone Python CLI.
- Deep Orthographic Control: Handles complex Persian typography including "heksare" detection and ZWNJ.
- Document Verification: Includes scripts to verify and sanitize DOCX and PDF files to prevent corruption and ensure font integrity.
- Bidi-Start Logic: Prevents layout scrambling in bilingual text by ensuring Persian sentences start with Persian words or invisible RLM characters.
- Bundled Fonts: Includes Vazirmatn and Lalezar fonts to ensure consistent visual rendering.
Related
- Project
- Project
- Project
- Project
- Project