About TafsirWeb

Most English tafsirs exist only as scanned PDFs — unsearchable, unindexed, and impossible to compare side by side. TafsirWeb digitizes them into a verse-keyed, permalinked corpus, and is honest about how far each work has been checked as that work proceeds.

14 tafsirs are digitized so far: Tafsir Ibn Kathir, Tafsir al-Jalalayn, Al-Mukhtasar, Maarif-ul-Quran, Tazkirul Quran, Tafsir al-Tustari, Lata'if al-Isharat, Ta'wilat al-Qur'an (2014 draft translation), Kashf al-Asrar, Asbab al-Nuzul, Tafsir al-Qurtubi, In the Shade of the Qur’an, The Meaning of the Qur’ān, Tafsir as-Sa‘di. A work shows at most one mark about its checking, beside its title. Across-checked, human or dual badge on a work means that most of its passages are checked at that level (one work is an exception, named below); no badge means the work is not yet checked. The project keeps a record of the checking of every passage, but a passage on the page carries no mark of its own.

Also available, when you turn on Show Urdu tafsirs in the reading settings: تفہیم القرآن. They are in Urdu and read right to left.

cross-checked means one of two things: two independent copies of the text agree, or the text is taken directly from its publisher's own file, the English text kept word for word, and the publisher is named as the source. Neither means a person has compared it with the printed book; that is whathuman and dual mean.

The Qur’an’s Arabic text and the English translations come fromTanzil. The default translation isSaheeh International; you can choose another in Reading settings. Each tafsir names its translator, editor and publisher in the “About this source” note beside the passage.

In the Shade of the Qur’an. For this work the badge refers to the surahs taken from the publisher's own files (surahs 1–39 and 78–114). Surahs 40–77 come from a single reading of the scan of the printed volumes.

Ta'wilat al-Qur'an. A translator's 2014 working draft, not a finished printed book. The editor's marks it still holds (such as "[?]", "[expand]" and "(ṣlʿm)") are printed here exactly as they stand in the draft.

No language model ever paraphrases, summarizes, or generates the tafsir itself. What you read is always a scholar's own digitized words.

This is a private, pre-alpha project. Its full architecture and current state are documented in IMPLEMENTATION.md and README.md in the project repository.