From an open question to an auditable research report
Taaweel is not a chatbot that answers in one shot. It runs a staged research method over the whole Quran corpus: it maps every occurrence, collects evidence in tiers, argues against its own conclusions, and writes a report where every claim points to a citation you can check.
One-shot AI answers are the wrong tool for Quranic research
A general-purpose chatbot gives a fluent answer and hides how it got there. For research that is a defect: you cannot see what was counted, which sources were consulted, which readings were rejected or why.
Taaweel replaces the single answer with a pipeline. Each stage is separate, saved and resumable, so a researcher can inspect the question as understood, the verses that were found, the commentary that was read, the objections that were raised and the rulings that were made on them.
- Transparent. Intermediate outputs are kept per session, not thrown away.
- Evidence first. Every
[E-n]citation in a report resolves to a row in the evidence ledger. - Self-critical. A red-team stage and a principles audit try to break the conclusion before you see it.
- Honest about status. Claims carry labels: mainstream, minority view, author position, provisional.
What Taaweel does
Six capabilities form one connected wheel. Each page goes into the detail of how it works and what it will and will not claim.
A staged method, not a single answer
Plan, map, collect, audit, compose. Each stage saved, resumable and reviewable.
Read moreEvidence and citationsEvery claim points to a ledger row
Evidence in five tiers, with every citation resolving to a ledger row.
Read moreRoot and word studiesSee every use of a root, then read what it means
Whole-corpus concordance, clustered by how a word is used, beside the lexicon.
Read moreAdversarial auditThe engine argues against itself first
A red team and 14 uṣūl principles try to break the conclusion first.
Read moreOntology explorerA concept graph that people govern
A concept graph where machine proposals wait for human adjudication.
Read moreMizan, companion appNumbers shown with their method and limits
24 interactive perspectives with named null models and stated limits.
Read moreHow a research run works
Refine the question
A short dialogue sharpens the question and fixes its scope before any compute is spent.
Map the corpus
Every occurrence of the key roots and terms is found and clustered into patterns of use.
Collect evidence
Verses, lexicon entries, tafsir positions and hadith are gathered into a tiered ledger.
Test the readings
The inherited reading and its alternatives are tried against linguistic, contextual and cross-verse tests.
Audit and attack
An adversarial stage raises objections; the audit checks the result against 14 principles.
Compose the report
Answer first, then the argument, with every claim cited and every limit stated.
Built for people who do this work
Quran researchers and scholars
Trace a concept across the entire corpus with the evidence and the objections laid out, not summarised away.
Educators and institutions
Give a seminar or a course a transparent workflow and saved sessions for review and assessment design.
Students of the language
See how roots behave across verses, then read the lexicon and tafsir entries that bear on a word.
Applied researchers
Psychology and medical teams can frame a research question against relevant verses with a documented trail.
What Taaweel will and will not claim
- It assists, it does not rule. Output supports scholarly work; it is not a fatwa and does not replace scholarly authority.
- It states uncertainty. Language models are probabilistic. Reports label provisional claims and keep the objections visible.
- Comparison is not identity. When a reading is compared with scientific findings, the report calls it a comparison.
- No borrowed credentials. Taaweel makes no claim of external certification.
Questions people ask
Is Taaweel available now?
Taaweel is in development and available for early access. Request access by email and we will share the current state, including what works end to end.
Which sources does it read?
The Quran corpus with root and token data, the Lane lexicon and classical dictionary entries, and tafsir databases including al-Ṭabarī and Ibn Kathīr. Corpora are opened read-only.
How is it different from asking a chatbot?
A chatbot answers once and does not show its work. Taaweel runs staged research, keeps every intermediate result, cites every claim to a ledger of evidence and runs an adversarial audit before writing the report.
Does Taaweel give religious rulings?
No. It is a research tool. It labels the epistemic status of each claim and is designed to support, not replace, qualified scholarship.
Can I export reports?
Yes. Reports can be exported as Markdown, HTML and DOCX.
What is Mizan?
Mizan is a companion application for transparent numerical analysis of the Quran text. It shows what was counted, how, and what the result does not establish.
See Taaweel run on your question
Tell us the question you are researching. We will walk through a run and what the evidence ledger and audit look like.