Hudson Legal Technologies
We work across legal AI — the corpus it has to be grounded in, the research into what is actually in that record, and the tools practitioners use every day. One standard runs through all of it: nothing is asserted that cannot be traced back to the text.
Documents in the corpus
Statutes, administrative rules and court rules
Iowa decisions
Citations mapped between those decisions
01 — Research
Building a research system means computing things about the law that nobody had computed before: every citation between every Iowa appellate decision, every act that has amended the Code, every rule an agency hangs off a statute. Our data briefs publish that work — original analysis of the Iowa record, answered from the whole of it rather than a sample, and written so a lawyer can check it.
One question per brief, with a point of view — not a filter panel handed to the reader.
How it was measured, what it misses, and the full data behind every figure.
Published with an as-of date and no drift. A refresh is a deliberate, reviewed re-export.
The fifty decisions Iowa's appellate courts cite most, counted from 475,375 citations across 76,725 decisions. The case they reach for most is not a constitutional landmark: it is In re P.L., a 2010 termination decision cited by 1,502 opinions. The full list, the chart, and how it was counted.
02 — Products
Flagship — research
A grounded research assistant for practitioners. Ask a question in plain language; it searches the corpus, reads the controlling text, and answers with citations that link to the source — each one verified before you see it. When the law is silent, it says so.

Responses are drawn from the effective text of the law — not from a model's recollection of it.
Every citation is checked against the source before the answer is delivered. Failures are surfaced, not smoothed over.
Statutes, administrative rules, court rules, and caselaw in one searchable system — semantic and keyword retrieval, fused and reranked.
The same grounded research runs in the browser, over a production MCP endpoint, and as an assistant that answers your email.
New — court filings
Research is half the day; the other half is paper. A Chrome extension for Iowa's EDMS that turns docket work into one click — preview a filing beside the docket, then download it named the way your office actually files things, straight from the court to you.
Read any filing in a panel beside the docket — no downloads-folder detour, no losing your place in the list.
Filings land under clean, consistent names from your own rules instead of the court's — one at a time, or the whole docket at once.
Documents move from the court straight to you. Hudson never receives, stores, or reads a filing — by construction, not by policy.
03 — Operating principles
Every response traces to effective, citable text and is checked before it is delivered. Where the source does not exist, the system says so.
The analysis that makes the products work is worth reading on its own — so it goes out as briefs and articles, with the method and the data attached.
We build for solo practitioners, small firms, and in-house teams — not only the institutions with procurement departments.
Our consulting practice draws on software we design, ship, and operate ourselves — the diagram comes after the work.
04 — What we do
Research
Original analysis of the Iowa record — which cases do the work, what changes, who regulates — published frozen, sourced, and open to being checked.
Read the latest briefProducts
Hudson Corpus in the browser, over MCP, and by email — and Hudson EDMSpro for court filings. The research is verified; the filings stay yours.
Explore the productsConsulting
Strategy, custom software, data, and applied AI for teams that need it built correctly — delivered against the same standard we hold our own products to.
Engage our teamHudson Corpus is live in beta. Ask a question you already know the answer to — then follow every citation to its source.