2010
CourtListener begins.
CourtListener starts as a free legal search project and grows into a major public source for federal and state opinions, APIs, alerts, and citation tools.
Why now
Public case law and practical AI finally arrived on the same timeline.
Freecase is possible because two long-running changes have converged: the public legal corpus is more accessible than ever, and search technology is now good enough to make millions of opinions usable without hiding the sources from the user.
The bottleneck was never whether the law was public.
Judicial opinions and statutes occupy a special place in American law. They are public legal materials, not creative works that a publisher can own in the ordinary sense. But public status is not the same as practical usability.
For centuries, binding case law into printed volumes, organizing it, checking it, and updating it as new case law emerged was a vastly expensive, time-consuming process. Just as technology has moved past the costs of pressing vinyl records, times have changed: what was once expensive and laborious can now be made freely available. The law is not music. It is a fundamental right of all citizens.
Public law timeline
The source material became dramatically easier to build on.
2013
Free Law Project forms.
The nonprofit behind CourtListener organizes around a broader mission: using technology and data to make the legal ecosystem more fair, useful, and competitive.
2018
Harvard's scan becomes searchable.
The Caselaw Access Project brings more than 6.5 million decisions, built from over 40 million scanned pages, into a researchable public dataset.
2024
The commercial-use cloud lifts.
As legacy restrictions around the Harvard dataset expired, public legal-data projects could work with a more complete public case-law corpus at bulk scale.
The important change is not a slogan. It is the existence of a large, inspectable, downloadable body of American case law that developers, researchers, lawyers, and public interest projects can actually build from.
Search timeline
At the same time, search technology became more useful for long legal texts.
2017-2020
Transformers changed language software.
Transformer models made it practical for software to compare, summarize, and sort long passages by more than exact keyword overlap.
2021-2025
Layered search became the consensus pattern.
Legal search research repeatedly points to a layered approach: exact wording first, broader concept matching where it helps, then careful ranking and testing.
Now
The costs and tools fit a public-interest product.
Open-source models, cheaper inference, strong databases, and better evaluation make it realistic for a small team to build a serious legal search system.
The Freecase bet
A shorter path from question to authority.
The goal is not to make legal research feel magical. The goal is to make the path from question to authority shorter, cheaper, and easier to audit.
Notes.
- Harvard Library Innovation Lab describes the Caselaw Access Project as more than 6.5 million decisions and explains that Harvard digitized over 40 million pages, representing more than 6.7 million cases and 360 years of U.S. legal history: Caselaw Access Project.
- Free Law Project describes its mission as using technology and data to make the legal ecosystem more equitable and competitive: Free Law Project.
- CourtListener provides free legal search and APIs for federal and state case law: CourtListener and its REST API documentation.