U.S. Federal and State Court Cases from the Caselaw Access Project

The Caselaw Access Project provides 6.7 million U.S. federal and state court decisions and judges' opinions. It includes nearly 40 million pages of legal text spanning the last 365 years, supplemented by over 900 thousand cases scraped from 479 courts via Court Listener. The dataset was created by common-pile and last updated on June 6, 2025.

Use Cases

• Train legal document classification models based on the corpus of court decisions and opinions.
• Perform temporal analysis of legal precedent evolution based on 365 years of coverage.
• Extract named entities and relationships for knowledge graph construction from judges' opinions.
• Conduct large-scale text mining for legal language patterns across federal and state jurisdictions.
• Benchmark information retrieval systems for legal case search based on the millions of pages of text.

Strengths

• Contains 6.7 million individual cases, providing substantial scale for analysis.
• Covers a 365-year time range, enabling long-term historical legal studies.
• Aggregates data from two major sources: the Caselaw Access Project and Court Listener.
• Includes text from nearly 40 million pages and over 900 thousand scraped cases from 479 courts.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Caselaw Access Project and Court Listener.

Collection Method

Aggregated from a wide variety of resources, including scraping from 479 courts.

Time Range

Last 365 years.

Freshness

Last updated 2025-06-06 03:51:35; freshness should be verified.

Geography

United States (federal and state courts).