Skip to main content

Independent research prototype. Not a U.S. government website; not legal advice. The complete disclaimer appears in the footer of every page.

RegLens-31
  • Overview
  • Search & browse
  • Extracted obligations
  • Statutory authority
  • Draft skeletons
  • Evaluation

Search & browse the corpus

Two utilities over the ingested corpus: full-text search across extracted claims, U.S. Code sections, CFR sections, and draft skeletons; and a part-by-section browser for the five ingested parts of Title 31.

Lexical search over extracted obligations, U.S. Code sections, CFR part sections, and draft skeletons. Exact-term matching against a precomputed index — no semantic ranking, no external service, and no model involvement.

Try an example query to explore the ingested corpus.

Indexed scope: extracted claims, U.S. Code sections, CFR sections, and draft skeletons.

Hierarchical navigation over the five ingested parts of 31 CFR (as of the pinned snapshot date), from part to section. Selecting a section opens the part text at that location. Paragraph-level drill-down is not built.

Independent personal research prototype. Not affiliated with, endorsed by, or representing the U.S. Department of the Treasury or any U.S. government agency (31 U.S.C. § 333). Not legal advice; no attorney-client relationship. Outputs are assistive only — verify every obligation against the primary source. Model-generated fields are labeled as such.

Data as of 2026-07-30 — a pre-computed static snapshot; no page depends on a live service. Source data: Federal Register (U.S. public domain). Extraction: qwen3:8b running locally at temperature 0.

Use-case framing: Treasury AI Use Case Inventory (U.S. Government work) — see About this demonstration · Static read API (JSON) · Source code & methodology ↗