Sunnah Search

A vector-database-backed AI search for discovering references across the Qur’an and the Sunnah—by meaning, not only by exact wording.

Ask in your own words; retrieval uses semantic embeddings so related verses, hadith, and commentary surfaces can align with your question’s intent.

What it is

Sunnah Search indexes Islamic source texts (and trustworthy reference material you choose to include) into a vector database. Each passage is represented as a high-dimensional embedding that captures meaning. When you search or chat, the system finds the most relevant passages before the model answers—grounding responses in actual citations rather than loose recall from training data alone.

How it works

  • Ingest & chunk — Texts are split into passages suitable for retrieval and attribution.
  • Embed — Each chunk is turned into a vector so “similar meaning” can be measured mathematically.
  • Retrieve — Your query is embedded and matched against the index (often with reranking for precision).
  • Answer with sources — The AI composes an answer while pointing to the Qur’anic āyāt, hadith, or other references the vectors surfaced.

Why semantic search

Classical study often moves between synonyms, translations, and related legal or theological themes. Keyword search misses many of those connections; semantic search helps bridge phrasing so you can explore themes—mercy, prayer, transactions, family—across revelation and prophetic practice with fewer dead ends.