Islamic historical sources have always existed — Ibn Sa'd, Al-Tabari, Ibn Kathir, Al-Dhahabi — but they've never been searchable with citations. Islamic Corpus fixes that. 141,566 chunks from 226 authenticated sources, queryable via API or natural language chat. Every answer cites the primary source — not Wikipedia, the actual text. Hybrid RRF retrieval with era filtering. Free tier: 50 queries/month, no credit card. Developer API plans showcased.
No reviews yetBe the first to leave a review for Islamic Corpus — The canon, made easy.
Maker
📌
Hey ProductHunt 👋
I built this because I kept running into the
same problem: ask any AI about Islamic history
and you get surface-level summaries with no
sources. Ask a search engine and you get
Wikipedia. But the actual primary sources —
Ibn Sa'd's biographical dictionary, Al-Tabari's
chronicle, Al-Dhahabi's 600 scholar biographies
— were sitting in public domain, untouched.
So I built a RAG corpus over 226 authenticated
Islamic sources spanning 632–1900 CE. The
retrieval uses hybrid RRF (vector + full-text
search with era filtering) so when you ask about
the Battle of Yarmouk, you get Al-Tabari's actual
account — not a chunk about hajj regulations that
happened to score well on cosine similarity.
Every response cites the primary source with
author and work title. Explorer mode gives you
the story. Research mode gives you full source
criticism — chain of transmission, where scholars
disagree, what Ibn Kathir said vs what western
academics argue.
Free tier is genuinely free — 100 queries/month,
no credit card. The dataset is also on HuggingFace
with 50+ downloads in the first 3 days.
Happy to answer any questions about the retrieval
architecture or the sources included.
GitHub: https://github.com/izzy19852/Isl...
HuggingFace: https://huggingface.co/datasets/...
Chat demo: https://islamiccorpus.com/chat
API docs: https://islamiccorpus.com/docs