yazinsai / quran-validator

What is quran-validator?

quran-validator is a lightweight TypeScript library that scans LLM-generated text for Quranic verses, validates them against a bundled authentic database with high accuracy, and auto-corrects subtle errors like missing diacritics or word swaps. It tackles AI hallucinations in religious content—where models misquote sacred text—via system prompts that make LLMs tag quotes parseably, plus post-processing for verification and fixes. Developers get zero-dependency validation, fuzzy matching across tiers (exact to normalized), and outputs like corrected text or issue reports.

Why is it gaining traction?

It stands out with ready-to-use LLM prompts (XML, Markdown, bracket formats) and untagged Arabic scanning, ensuring high accuracy even for real-world messy outputs. No setup hassle: npm install, process text, get validated results with confidence scores. The full 6,236-verse Uthmani database and TypeScript types make it drop-in reliable for Arabic validation workflows.

Who should use this?

AI builders creating Islamic chatbots, Quran study apps, or fatwa generators powered by LLMs. Content platforms needing to validate religious explanations before publishing. Devs integrating generative AI into Arabic/Islamic sites where quote fidelity is paramount.

Verdict

Use it now if accuracy in religious AI output matters—it's battle-tested via LLM benchmarks showing real fixes. Modest 121 stars and 1.0% credibility score reflect early maturity, but strong docs, tests, and npm package make it viable for production with minimal risk.

(198 words)