Getting started
How to use the engine
Hadith Compass is a research tool, not a verdict machine. Follow these five steps to load data, analyze narrations, and compare ethical readings across frameworks.
Ingest a collection
Go to Ingest and select one or more canonical collections (Bukhari, Muslim, etc.). Choose how many narrations to load per collection, preview the data, then confirm. Already-imported narrations are skipped automatically.
Start small — 50–100 narrations per collection is enough to explore the engine.
Browse the Library
Open the Library to search and filter narrations by collection, grade, topic, and target audience. Each card shows how many frameworks have analyzed it so far.
Use the search bar for keywords, or the dropdowns to narrow by topic or target audience.
Select & analyze
Select individual hadiths (or use 'Select unanalyzed only') and click Batch analyze. The engine runs AI classification and ethical scoring under every active framework. Each hadith becomes 3+ separate analyses.
Each hadith analyzed costs roughly 3–4 AI calls (one for classification, one per framework). Watch the progress panel for status.
Review Analytics
Switch to Analytics to see aggregate scores, topic distributions, and cross-framework comparisons. Understand how the same corpus reads differently under UDHR, Traditionalist, and Reformist lenses.
The radar charts show divergence between frameworks — that divergence is the insight, not any single score.
Read the Methodology
For the full picture of how classification, scoring, and safeguards work, visit the Methodology page. Every score is framework-relative and carries confidence values and alternative interpretations.
No score is a moral verdict. The engine is descriptive, not prescriptive.
What this engine does — and does not do
- · It describes narrations through explicit ethical frameworks, never claiming objectivity.
- · It scores each hadith independently under each framework, so you can compare readings side by side.
- · It surfaces alternative interpretations and confidence values on every score.
- · It does not issue fatwas, rank religions, or replace scholarly expertise.
- · It does not claim any single score is "the truth." Divergence between frameworks is the point.