Caselaw Access Project
The Caselaw Access Project (CAP) scanned the entirety of Harvard Law School Library's physical collection of American case law and turned it into machine-readable data. It includes all official, book-published state and federal United States case law through 2020: every volume or case designated as an official report of decisions by a court within the United States.
In total it covers 6.9 million decisions reaching back to the 1600s. Earlier metered access agreements have expired, so there is no longer any restriction on how much of the data you can pull, and a free bulk API is available for large downloads.
The court opinions themselves are in the public domain, and the digitisation work is released under CC0, a public domain dedication that waives all rights. A full mirror is also published on HuggingFace, which is often the quickest way to load the corpus into a pipeline.
Related sources
AustLII
A free resource of medium-neutral case law and unreported judgments for all Australian jurisdictions, covering the whole country since 1995.
Awesome LegalTech
A curated list of legal technology resources: open-source platforms, AI models, companies, datasets, and tools spanning the global legal ecosystem. Useful for tracking new entrants in a fast-moving space.
BAILII (British and Irish Legal Information Institute)
Free access to British and Irish primary legal materials, covering UK and Ireland case law and legislation. Alongside the National Archives, one of the main free sources for reading UK judgments.
Cambridge Law Corpus
A research dataset of more than 250,000 UK court cases, mostly from the 21st century but with some reaching back to the 16th century. Built by the University of Cambridge for legal natural language processing work, it is available under restricted access for research use.