State Statutes in Bulk: Which States Publish XML, JSON or an API?

Virginia's law site hands you a list of every title in the Code of Virginia as JSON with one request and no key. Its library page also says its CSV files have fields separated by vertical bars. The files we opened use commas. Details like that decide whether a state's "data" page saves you a week or costs you one.

Short answer: few state legislatures publish their statutes as data. Of 19 states read for this guide, six offer statute text in a structured format: Virginia (JSON, XML and CSV, no key), Utah (XML), New York (JSON, free key), Michigan (one XML file per chapter), Iowa (one XML file per chapter) and California (a database dump with XML text). Nine more post fixed chapter or title files, such as HTML, PDF or Word. Two give you documents or software only. Two offer data about bills, not statute text. This is a named sample of 19 states.

Written and checked on 12 October 2026. This is general information, not legal advice; check the cited statute or opinion before relying on it.

TL;DR

• "Bulk" means four different things: a data API, a structured file, a copy of rendered pages, or documents meant for people. Most states offer the last two.
• The six structured sources each have a catch: a key (New York), a once-a-day request recommendation (Utah), a trailing-slash redirect and a wrong format note (Virginia), a database to load (California), file dates that vary widely by chapter (Michigan), or XML links with no official or unofficial label on the page (Iowa).
• States label their own copies. Utah calls its XML official. Illinois, Oregon, North Carolina and South Carolina say theirs is not. Store the label with the text.
• Bill data and statute data are different things. Maryland and Washington publish good data about bills. The codified law comes from a different source.
• Nothing here is a census. Nineteen states were read; eight were read through a web crawl because the site did not answer from our network.

Official text, as of 12 October 2026

Citation: Utah Legislature, Developer Tools page, "Utah Code XML" entry. Read through a web crawl on 12 October 2026, because le.utah.gov did not answer from our network.

What it says: "The code provisions available above constitute the current and official electronic record of the Utah Code."

Official link: le.utah.gov/data/developer.htm

What changed in 2025 and 2026: the "current through" dates on the pages we read run from 1 January 2025 (Connecticut's revised-to date, with a 2026 Supplement for later changes) to 1 October 2026 (Maine and Wisconsin). Illinois's FTP copy sits in between at 21 November 2025. That spread is the main reason to record a read date for every state.

Four things people call bulk data

The phrase covers four different products, and they need four different pipelines.

1. A data API. You send a request and get structured text back. Virginia and New York do this for statutes.
2. A structured file. You download XML, CSV or a database dump and parse it yourself. Utah, Michigan, Iowa and California fit here.
3. A copy of rendered pages. Fixed URLs return one HTML, PDF or Word file per chapter or title. South Carolina, Maine and Illinois are examples. You parse the layout.
4. Documents for people. PDFs, e-books or a desktop program. Florida and Colorado are examples. You can read the law there. Getting it into a database is your job.

What we checked, and how

We read the official statute or data pages of 19 states on 12 October 2026. Where the site answered from our network, we ran the request and recorded the HTTP status, content type and size, or read the file's zip directory with range requests. That covers Virginia, California, South Carolina, Maine, Texas, Delaware, Iowa, Florida, Maryland, Washington and Colorado.

Eight sites did not answer from our network or returned a block page: Utah, New York, Illinois, Oregon, Michigan, North Carolina, Connecticut and Wisconsin. We read those through a web crawl of the official page. A crawl read tells you what the page says. It does not prove a download works, so those rows are marked and no request is claimed for them.

Not covered: every state not in the tables. We also tried Massachusetts, Ohio, Pennsylvania, New Hampshire, Nevada, Alaska, Vermont and Idaho, and could not read their pages. That says nothing about what they offer. For the wider method, see how to run a 50-state survey.

The matrix

Structured statute text

Fixed chapter or title files

Documents or software only

Bill data only

What the rows show

Virginia is the most developer-friendly source we read. Its developers page says the web services are "available in both JSON and XML". Section text comes back with the title, chapter and section in each record. The titles list repeated three title numbers (8.2, 8.2A and 8.4A) when we ran it, so key your records on more than the title number.

California is the deepest dump, and the heaviest. The current-session file, pubinfo_2025.zip, was 1,288,568,136 bytes (about 1.29 GB) when we checked. It holds the code tables. The daily archives do not, and the readme says so in a single line. The scraping guide covers the same files at a higher level.

Utah and New York trade ease for rules. Utah's developer page lists its endpoints with no key and asks you to cache and, for the code, check no more than daily. New York needs a free key, and its date option is the clearest as-of tool we found in a state API.

Michigan's XML is real law data with uneven freshness. The directory listing shows a modified time per chapter, and those times range widely. Chapter 257 showed 23 September 2026 and Chapter 119 showed 24 September 2021. A file's date shows when the file changed. It does not show that the chapter is complete through the latest act, so compare against the HTML pages before you rely on one.

South Carolina is the only state here whose page states reuse in plain words. It covers code text, numbering, history and some notes, and it also says the web version is not official. For other states, the reuse question is open. See Can you use state statutes in a commercial product?.

Iowa offers XML per chapter, and an e-book that has not changed since 2022. The chapter listing for Iowa Code 2026 has an XML link beside each PDF and RTF link. The Chapter 39 XML file served a last-modified date of 10 July 2026. The EPUB e-book served 10 January 2022, while the first volume PDF showed 22 December 2025. A download link on a page does not mean the file is current.

Florida's download is a program. The zip is about 251 MB. Its file list shows an installer, supporting software and library files with the extension .nxt, the largest about 242 MB. We did not open those files.

For the bill side, the LegiScan API guide and the Open States API guide cover aggregators that pull bill data from many legislatures.

Requests we ran

Virginia, a section of the Code as JSON (12 October 2026). Without the trailing slash the server answers 301, so use -L.

curl -sL https://law.lis.virginia.gov/api/CoVSectionsGetSectionDetailsJson/11-1/

The response was HTTP 200, application/json, 1,591 bytes. Trimmed:

{"TitleNumber":"11","TitleName":"Contracts","ChapterList":[{"ChapterNum":"1","ChapterName":"General Provisions","SectionNumber":"11-1","SectionTitle":"Certain contracts void as to creditors and purchasers unless in writing; law governing validity of contracts creating security interests","Body":"<p>Every contract, not in writing, made in respect to real estate or goods and chattels in consideration of marriage, ..."}]}

Virginia's library page says its CSV files are "delimited by vertical bars". The Title 11 file we downloaded (41,535 bytes, 39 data rows, 15 columns) was comma-separated with quoted fields, and the body column holds HTML.

California, the size and type of the yearly file before you commit to it:

curl -sI https://downloads.leginfo.legislature.ca.gov/pubinfo_2025.zip

HTTP/1.1 200 OK
Last-Modified: Mon, 12 Oct 2026 04:26:09 GMT
Content-Length: 1288568136
Content-Type: application/zip

We read the zip's file list with range requests instead of downloading it. It includes CODES_TBL.dat and LAW_SECTION_TBL.dat (43,276,148 bytes uncompressed). The first record in that table is Penal Code section 4027.5, with the section text in a .lob file that begins <caml:Content xmlns:caml="http://lc.ca.gov/legalservices/schemas/caml.1#">. The pubinfo_daily_Sun.zip file (1,158,201,706 bytes) lists 14 tables, and none are law or code tables. The readme describes the daily files as "all current session data except Code tables."

Washington, bills that touch one RCW section (this returns bill data; the RCW text itself is on web pages):

curl -s "https://wslwebservices.leg.wa.gov/RCWCiteAffectedService.asmx/GetLegislationAffectingRcwCite?biennium=2025-26&rcwCite=9.41.010"

The response was HTTP 200, text/xml; charset=utf-8, 7,751 bytes, with 16 bill entries. The first was HB 1281.

Maryland, bill data as JSON: https://mgaleg.maryland.gov/2026rs/misc/billsmasterlist/legislation.json returned HTTP 200, application/json, 8,202,805 bytes, a list of 2,677 bill records on 12 October 2026. The first was HB0001.

Utah, New York and Michigan are documented in the table but not run. Utah's and New York's hosts did not answer from our network, and New York needs a key. Michigan returned 403 to a script. Their URL patterns come from the states' own pages: glen.le.utah.gov/code/3-1-1/ for a Utah section, /api/3/laws/{lawId} under legislation.nysenate.gov for a New York law, and legislature.mi.gov/documents/mcl/ for the Michigan chapter files.

Official copy or convenience copy

States label their own downloads, and the labels differ in a way a parser cannot see.

• Utah calls its code XML "official".
• Delaware says its site "is considered an official version".
• Iowa marks its volume PDFs official and its chapter PDFs and RTF unofficial, and we saw no label on its XML links. Its page also describes authentication of official PDFs under the Uniform Electronic Legal Material Act (UELMA), with an implementation date of 9 January 2020.
• Illinois, Oregon, North Carolina and South Carolina say their online copies are not the official text.

Delaware's page says UELMA took effect in the state on 21 October 2014. We did not check the enactment list for other states.

Store the label, the URL and the read date with every section. For the concept behind it, see official vs unofficial legal sources.

Which shape fits which job

Statute text changes less often than bills, so the two usually need different refresh schedules. For what happens when no bulk path exists, see what breaks when you scrape state legislature sites. For regulations, which are a separate source, see state regulations API.

What goes wrong

• "Downloadable" is treated as "official." Check each label, as in the section above.
• Bills are loaded instead of statutes. The Maryland and Washington data is about bills.
• A copy is older than it looks. Illinois generally updates in the fall, Oregon by edition, and Iowa's EPUB was last modified in January 2022.
• A readme goes unread. California's daily files leave out code tables, and large text sits in separate .lob files.
• The page and the file disagree. Virginia's page says vertical bars and the file uses commas. South Carolina's page source still holds a commented-out June 2017 update notice, while the visible disclaimer says the Code is current through the 2025 Session.
• Rate guidance is ignored. Utah recommends a daily check of its code. Michigan limits automated requests to one per second.
• "No link found" becomes "none exists." A state we did not read may publish more than any row here.

Questions to ask a vendor or a state office: what is official, how often does it refresh, what are the terms, and is there an as-of date. For the types of law behind these sources, see types of primary law in the US.

A short history

On 30 May 2013, Waldo Jaquith of the State Decoded project wrote that "like most states, New Jersey doesn’t provide bulk downloads." The post reported that Public.Resource.Org had put five state codes online as bulk data: Arkansas, Colorado, Georgia, Idaho and Mississippi, joining Maryland and Washington D.C. Read it at The State Decoded. We did not check what those states offer now.

The District of Columbia, which is not a state and is outside the table, has a public GitHub repository, DCCouncil/law-xml, described as "DC Law - statutes and code - in xml format." GitHub's record shows a push on 10 October 2026 and no licence on the repository. We did not read its terms or check whether it is the official text.

Sources read for this post

All read on 12 October 2026. Run means we sent the request. Crawl means the page was read through a web crawl.

• Virginia: law.lis.virginia.gov (home, developers, law library and the API and CSV paths above). Run.
• Utah: le.utah.gov/data/developer.htm. Crawl.
• New York: legislation.nysenate.gov/static/docs/html/index.html and laws.html. Crawl.
• Michigan: legislature.mi.gov/documents/mcl/, the Automated Data Collection Policy page and the FAQ. Crawl.
• California: downloads.leginfo.legislature.ca.gov and pubinfo_Readme.pdf; Government Code section 10248.5 (crawl).
• South Carolina, Maine, Texas (tcss.legis.texas.gov and statutes.capitol.texas.gov), Delaware, Iowa (the Iowa Code, chapter listing, official and unofficial editions, e-reader and certified-copy pages), Florida, Maryland, Washington, Colorado: official pages and files named above. Run.
• District of Columbia: github.com/DCCouncil/law-xml, read through the GitHub API.
• Illinois: ilga.gov/ftp/ILCS/aReadMe/aReadMe.txt. North Carolina: ncleg.gov chapter files, statutes page and disclaimer. Oregon: oregonlegislature.gov ORS page. Wisconsin: docs.legis.wisconsin.gov/statutes. Connecticut: cga.ct.gov General Statutes pages. Crawl.

This guide is part of US Law Data: The Complete Guide, a map of where US law comes from and how to use it.

FAQ

Which states publish their statutes as XML?

Among the 19 states read, Utah serves the Utah Code as XML, Michigan and Iowa post one XML file per chapter, Virginia offers an XML version of its web service, and California's text sits in XML inside a database dump. This is a sample, not a census.

Does any state have a statutes API?

Virginia has a REST web service with no key, in JSON and XML, and New York has a REST API that needs a free key. Utah serves XML at fixed endpoints that work like an API. Washington and Maryland offer data about bills only.

Can I download all of a state's statutes at once?

California offers a yearly database file. Michigan, Iowa, Virginia, South Carolina and Maine offer files you can fetch chapter by chapter or title by title. Florida offers a program, not a data file. For most other states in this sample, you fetch pages.

Is a state's downloadable copy the official law?

It depends on the state. Utah calls its XML official, and Delaware says its site is considered an official version. Illinois, Oregon, North Carolina and South Carolina say their web copies are not official. Check the notice on each source and store it.

How often are state statute downloads updated?

It varies. California's directory listing showed its current-session file modified on 11 October 2026. Illinois updates its FTP copy generally once a year, in the fall. Oregon updates by edition, and its next edition is due online in early 2028.

Is it legal to scrape a state's statutes?

Copyright, site terms and server load are separate questions, and they differ by state. Michigan and South Carolina post explicit rules. Read the terms, honor robots files and rate limits, and ask a lawyer if your product depends on it. See the licences guide and the scraping guide.

What if my state is not on the list?

Read the legislature's developer or download page, its readme, and its terms. Look for a directory listing at the paths the site uses for documents, and write down what you find, with the date.

For the US statutes, regulations and court rules these laws sit in, see the US primary law API.

Priyansh Khodiyar

Co-Founder & CTO

Priyansh leads engineering and AI at Vaquill AI: the pipelines that pull statutes, regulations and court rules from every US jurisdiction's official publisher, and the REST API, MCP server and open dataset that serve them.