Contact · soon
All work
01 / Model Context Protocol server

Legal Research MCP

A 45-tool research server that lets Claude search, read, and cite New York and federal law — backed by 5.1 million decisions and a citator built from their own text.

Client
A New York law firm
Role
Architecture, data engineering, build
Year
2026
Status
In production · private
Live
Private deployment
Claude · legal research connector
Is Pell still the standard for arbitrary-and-capricious review? Pin cite, and how often the Court of Appeals still relies on it.
citing_references("34 NY2d 222") ✓ 0.2s
get_ny_case("34 NY2d 222") ✓
Yes. Matter of Pell v Board of Educ., 34 NY2d 222, 231 (1974): arbitrary action is “without sound basis in reason and . . . taken without regard to the facts.” The local citator finds 5,202 citing decisions — 247 from the Court of Appeals — and no formal negative treatment. Most recent Court of Appeals use: Matter of Mantilla v NYC Dept. of Hous. Preserv. & Dev., 2025 NY Slip Op 07079.

45 tools · 12 sources

  • NY Court of Appeals & App. Div.
  • NY slip & motion decisions
  • Federal courts
  • NY Consolidated Laws
  • NYS bills
  • U.S. Code
  • Code of Federal Regulations
  • NYCRR
  • Federal Register
  • Regulatory dockets
  • Agency opinions
  • Court dockets
  • Verdicts
  • Practice treatises
  • Contract exemplars
  • Citator
Citator data as of Sep 24, 2026
45research tools exposed to Claude
5.1Mcourt decisions held in full text
41.7Mcitation links in the citator graph
600+automated tests

The brief

Why it needed to exist.

General-purpose AI models invent citations. Lawyers need a research tool that answers with real, checkable authority — and says plainly when it can't find something, instead of guessing.

The production

What we built.

The engine underneath everything in this act. It connects Claude to primary law through the Model Context Protocol — case law, statutes, bills, regulations, dockets, agency opinions, verdicts, and treatises — and every result carries the case name, court, date, official and parallel citations, and a link back to the source.

Where the free public APIs stop, it keeps its own corpus: the New York Official Reports, nearly a million New York slip and motion decisions refreshed nightly, and roughly 3.3 million federal decisions de-duplicated across three independent sources.

  1. 01Forty-five tools, twelve sources

    Public legal APIs — CourtListener, NYS OpenLegislation, eCFR, the Federal Register, GovInfo, and Regulations.gov — alongside local corpora for New York and federal decisions, the NYCRR, agency opinions, verdicts, and licensed treatises.

  2. 02A citator built from scratch

    A citation graph extracted from the decisions' own text: who cites a case (with the citing sentence), negative treatment, and direct history. Accuracy was measured by hand rather than assumed, and it is always labeled as not Shepard's or KeyCite.

  3. 03Connects like any Claude connector

    Standards-based OAuth 2.1 sign-in, so a colleague adds it to Claude with a URL and an account — no local setup.

  4. 04A good API citizen

    Respects every upstream provider's rate limits, caches aggressively, and identifies itself honestly.

  5. 05Keyword search you can check

    Full-text indexes ranked by relevance — no embeddings, so every hit is a literal match a lawyer can verify on the page.

  6. 06Always current

    Corpora and the citator update nightly and incrementally, without anyone touching them.

Under the hood

How it’s made.

  • RuntimePython with the MCP Python SDK; local and remote transports
  • DataSQLite full-text search across seven indexes, ranked by BM25
  • CitatorCitation extraction with eyecite plus a custom canonicalizer; incremental nightly updates
  • AccessOAuth 2.1 with PKCE and dynamic client registration for claude.ai
  • Quality600+ automated tests and guarded, repeatable deploys

Stack: Python · Model Context Protocol · SQLite full-text search · OAuth 2.1

Next productionCite Check for Word