LOCUS

LOCUS: A large-scale corpus of U.S. local ordinances for legal AI research

Researchers introduce LOCUS, a comprehensive machine-readable corpus of U.S. municipal and county ordinance codes covering 9,239 jurisdictions, with a county-harmonized access layer spanning 2,309 of 3,144 U.S. counties. The corpus was assembled using OCR to handle diverse document formats previously locked in vendor platforms, and is released on HuggingFace alongside ModernBERT-based classifiers for analyzing local law along dimensions like opacity and paternalism. The work addresses a significant gap in legal AI training data, as local ordinances govern large swaths of everyday regulation but have been absent from existing corpora.