بِسْمِ ٱللَّٰهِ ٱلرَّحْمَٰنِ ٱلرَّحِيمِ
PROVENANCE & DEFENSIBILITY
TheoAI runs on the Islamic Primary Source Corpus (IPSC) v3.4, deployed on Microsoft Azure AI Search across 13 customer-facing alias indexes. This page summarises the corpus’s provenance, the IP posture, and the audit transparency that together make IPSC the only defensible AI hadith corpus in market.
Authoritative one-line summary
IPSC is an applied-AI / data-engineering work product grounded in classical rijāl methodology. It is not classical mujtahid scholarship. Canonical external citation: “AI-assisted hadith corpus, structurally validated against documented teacher-student relationships, with an open scholar-collaboration program.”
Source: corpus-v3/manifest.json _provenanceDisclosure (v3.11
methodology recalibration, 2026-05-02). For the full block, see
ipsc.theogrid.ai/provenance.
449,285
Hadith records
v3.4 deployed; v3.26 staged adds 8,241 more from 6 new collections
27,118
Narrators (NRS)
Anchored to Ibn Ḥajar Taqrīb al-Tahdhīb; 12-tier reliability scoring
11
Vector indexes (3,072-dim)
OpenAI text-embedding-3-large; matn x 3, narrator x 3, defect x 2, term x 3
~$420
Cumulative LLM spend
v3.5 → v3.26; per-release budget detail in /changelog
US patent applications filed for the matn-criticism pipeline (two-pass architecture: deterministic string-op + multi-tier LLM reasoning), the NRS synthesis method (classical Taqrīb anchor + multi-source reconciliation), the semantic cluster + chain-matn conflict detection technique, and the classical-dimension computational approximations (shudhūdh, idrāj, maqlub, muḍṭarib).
IPSC™, MindHYVE™, Eve-Theology™. Common-law in use; USPTO filing in preparation.
Specific implementation details (LLM prompts, threshold values, curated reference databases, orchestration logic) protected as trade secrets under DTSA and Delaware UTSA. Methodology disclosed; specifics gated behind reviewer agreement.
EU Database Directive 96/9/EC sui generis rights on the compiled corpus, NRS database, matn clusters, and ilal cross-links.
Full IP statement at licensing@mindhyve.io;
IP-NOTICE.md in the corpus repository is the canonical document.
Most AI products in the Islamic-content space sit in one of two camps: “trust the model” (no provenance, no QA discipline) or “we have classical scholars” (no scale, no documented practice). TheoAI sits in the third, harder, position: AI-assisted, structurally validated, with a documented continuous quality cycle, recurring independent review, and an open scholar-collaboration program — every layer published.
Recurring
External-examiner cadence
Independent reviewers pressure-test framing on a recurring cycle. Most recent cycle 2026-05-02; recalibration items shipped in v3.11.
_provenanceDisclosure
Manifest block on every release
AI-involvement scope ships with every record so consumers can cite the disclosure alongside any IPSC claim.
~98K
Scholar-collaboration program
Open scholar-collaboration program; the number is the size of the program, not a backlog. Working muḥaddithīn engaged on a continuous basis.
33/33 + 7/7
Hard regression gates
33-test regression suite + 7-test cross-field consistency check; both green through every release v3.5 → v3.26.
Technical disclosure
The full _provenanceDisclosure block from the corpus manifest, in scholar-facing register.
Audit practice
The four pillars of IPSC quality, the recalibration timeline, and the open scholar-collaboration program.
Release history
Full v3.0 → v3.26 timeline with deployment status and per-release highlights.
Live API
3-tier API access (Public / Research / Scholar) on Azure AI Search infrastructure. Commercial access via licensing.
بِسْمِ ٱللَّٰهِ ٱلرَّحْمَٰنِ ٱلرَّحِيمِ
I'm here. Ask me anything about Islam. Every source cited. Every chain verified.
Ask about any hadith, ruling, or scholar…449,415 graded hadith · 27,115 verified narrators · 95.2% Bukhari convergence