Jurisdiction-scoped corpus
Official Jordanian documents are ingested, normalized, and versioned provision by provision. Retrieval is scoped to active jurisdictions — Jordan today, MENA as it is onboarded.
Jordan-first · jurisdiction-aware · source-verified
Authoritative Jordanian legal knowledge, engineered for retrieval.
LKC turns validated legal documents into a searchable public corpus — then answers questions grounded only in retrieved provisions, with citable authority and an auditable trail.
Built on four guarantees
Citation-verified answers
Every answer traces to source provisions ranked by authority tier. Verification confirms the cited passages ground the output — no retrieved context, no claim.
Auditable skill execution
Workflows run as versioned, sandboxed skills with approval and security gates. Every run is recorded append-only, and external skill use is governed by a provenance and license review.
Security by construction
Row-level security plus a retrieval view that structurally excludes private and matter data. The public surface can never mix with tenant data.
How a question becomes an answer
- Ingest — validated source documents are hashed, parsed, and versioned.
- Normalize — Arabic legal text is keyword-normalized so retrieval matches reliably.
- Retrieve — hybrid search over the public corpus, jurisdiction-scoped.
- Verify — the answer is grounded in and cited against the retrieved provisions.
Summary-verified API
GET /api/v1/skills
POST /api/v1/skills/{id}/execute
POST /api/v1/legal/answerKey-scoped endpoints, jurisdiction and visibility enforcement, and rate limiting. Machine clients authenticate with scoped API keys — secrets never reach the browser.
WakeelyPro
Lawyer mode — matter-aware workflows, full citation trails, human-review gates.
Mokhamen
Citizen mode — plain-language answers to everyday legal questions, grounded in the same public corpus.