Multi-substrate retrieval fusion
The engine blends passage search, citation-chain search, holdings, maxims, documents, parties, facts, and evidence into one research path instead of relying on a single result list.
QI Methodology
LAW is built around QI: a deterministic, citation-grounded legal intelligence engine. The method is simple in principle and strict in practice: retrieve from primary sources, verify the authority path, audit the filing risk, and hold external action behind review gates.
1,900,000+
Primary-Source Passages
995
Source Texts
530,080
Citation References
87,085
Cross-References
Verification Spine
The hosted workspace is designed to keep legal work close to the record. It does not treat fluent text as authority. The source path, treatment status, forum fit, and filing posture remain visible until the work is ready for review.
Retrieve
QI searches the allowlisted corpus, citation references, holdings, maxims, and matter record before any legal assertion enters a workspace.
Verify
The verifier separates citation existence, quote support, source fit, negative treatment, forum authority, and unresolved review needs.
Audit
Packets are checked against jurisdiction, standing, service, formatting, citations, exhibits, and completeness before filing reliance.
Approve
Client-facing, served, or filed work remains locked behind review gates. Research can move quickly; external action still requires authorization.
The Engine
Some layers are visible in the hosted product today. Other layers are engine-side methodology that powers retrieval, doctrine mapping, and gap research. Public claims are separated accordingly.
The engine blends passage search, citation-chain search, holdings, maxims, documents, parties, facts, and evidence into one research path instead of relying on a single result list.
QI studies primary sources and expands doctrine maps over time. New understanding is promoted only through controlled research and verification paths.
Matter facts, evidence, entities, authorities, and procedural posture are represented as a graph so legal conclusions can be traced back to their factual and legal anchors.
Authority is weighted by court, jurisdiction, treatment, and procedural context. A strong source in one forum may be weak or inapposite in another.
QI Research can study judge, forum, and doctrine patterns. The hosted platform presents those ideas as research methodology unless a specific workflow is verified in code.
When authority is missing, conflicting, or unsupported, the engine queues the gap for research instead of filling it with unsupported language.
Verified Now
This is the platform boundary: these capabilities are safe to feature because they are backed by code paths, public constants, or implemented workflow surfaces.
Research Boundary
LAW publishes research in public corpus theory, doctrine mapping, and structural legal analysis. Those papers may report experimental measurements. The product does not present those measurements as guaranteed outcomes, litigation predictions, or automatic filing power.
Doctrine-cluster experiments, pre-consensus signal studies, dormant rights research, and manufactured precedent maps belong to the QI Research series unless a hosted workflow separately proves the behavior.
Public capability language is limited to the corpus, verification spine, filing packet workflows, review gates, and supported e-submission portals that are backed by this codebase.
Research notes, drafts, citations, and packets are retrieved and verified from primary sources - not generated as final authority - and remain subject to user and professional review.
Institution-Grade Control
QI can accelerate retrieval, mapping, drafting, and audit preparation. It does not remove the duty to verify. LAW keeps the source record, review state, and filing authorization in the workflow so unsupported work is stopped before it leaves the workspace.