quran-semantic-search open-sources Chinese Quran RAG corpus

The core value here is not the UI, it’s the curated parallel corpus: a narrow, high-consistency dataset that is actually useful for retrieval, comparison, and instruction tuning in Chinese. That makes it more interesting as a domain data asset than as a general-purpose app.

• –Five aligned translations give you a built-in comparison set for semantic retrieval and answer-grounding experiments
• –ShareGPT and Alpaca exports reduce friction for LLaMA-Factory, Axolotl, and similar training stacks
• –The static SSR front end is a nice demo layer, but the dataset is the durable part of the release
• –This is strong niche infrastructure for multilingual RAG, but it is not broad enough to matter outside religious/textual domains
• –Licensing and provenance matter here; translation corpora can be useful without being free of copyright or attribution constraints