About the data

Local Law Explorer is a free, non-commercial tool for reading the city and county ordinances that shape everyday life. It is built entirely on LOCUS-v1, an open corpus of U.S. local law compiled by Denis Peskoff, Joe Barrow, Christopher Vu, and Diag Davenport. This site does not compile any of the data itself — it is a reader for their work, and all credit for the underlying corpus is theirs.

Local ordinances are the most fragmented body of law in the United States. There is no central index, and every town's code is published in its own place and format, which makes them nearly impossible to compare. LOCUS gathered them into one uniform table: 2,211,516 provisions from 1,911 cities and 376 counties across all 50 states, each labeled by topic and scored on several model dimensions. This site renders a detailed page for every one of them.

Where the text comes from

The LOCUS project gathered municipal and county codes published across many vendor platforms, ran optical character recognition (OCR) to turn scanned and exported documents into text, then split each code into individual provisions. Machine-learning classifiers labeled every provision by topic and whether it states a substantive rule. We keep only the substantive provisions and present them grouped by topic.

What the labels mean

Topics (Zoning, Nuisance, Buildings, Business, Other) are model predictions, not official categories. The three lenses on each jurisdiction page (everyday rules, starting a business, renting and property) are our own keyword groupings layered on top of those topics to make the text easier to navigate. The “densely worded” flag reflects an opacity score from the LOCUS models. Treat all of these as helpful signposts, not authoritative classifications.

How the comparisons work

Each jurisdiction page opens with a short portrait of how that place governs. LOCUS scores every provision on four model dimensions, including how densely it is written, how much it regulates personal conduct, and how much it leaves to officials' judgment. We average those scores per jurisdiction and rank each place against the roughly 2,300 cities and counties in the corpus, so a statement like “writes its laws more plainly than 70% of towns” is a percentile, not a verdict. These are machine estimates and we label them as such. We omit one dimension (problem salience) from the portrait because its meaning is not settled enough to describe in plain words.

The Everyday questions section matches common questions to the ordinances whose titles mention them, and shows the full text of each. A keyword match is not a yes-or-no answer; it points you to the rule to read. The Notable rules section uses a simple text heuristic to surface distinctive subjects (animals, parades, curfews, and the like). It matches words, not meaning, so it is a curiosity rather than a survey.

Important limitations

• This is not legal advice and is not a substitute for a lawyer.
• Text is OCR'd and may contain scanning errors; labels are automated and unaudited.
• There are no effective dates or amendment history. A provision shown here may be outdated, amended, or repealed.
• Coverage is partial. For each county the corpus keeps one representative code, so your specific town may not appear.
• It does not resolve which level of government (state, county, or city) controls a given question.
• Always verify against the official municipal code.

Credit & license

LOCUS-v1 is used under the Creative Commons Attribution-NonCommercial 4.0 license. This tool is non-commercial in keeping with that license.

@article{peskoff2026freeing,
  title   = {Freeing the Law with LOCUS: A Local Ordinance Corpus for the United States},
  author  = {Peskoff, Denis and Barrow, Joe and Vu, Christopher and Davenport, Diag},
  journal = {arXiv preprint arXiv:2606.19334},
  year    = {2026}
}

Dataset: huggingface.co/datasets/LocalLaws/LOCUS-v1 · Paper: arXiv:2606.19334