I spent 6 years indexing Indian court cases from the Supreme Court, all 25 High Courts, and 14 Tribunals. Sharing because I haven’t seen a structured Indian legal dataset at this scale anywhere.<\/p>\n
What’s in it:<\/p>\n
– 20M+ cases with pdf, structured metadata (court, bench, date, parties, sections cited, acts referenced, case type, headnotes)<\/p>\n
– Citation graph across the full corpus (which case cites, follows, distinguishes, or overrules which)<\/p>\n
– 23,122 Indian Acts and Statutes (Central, State, Regulatory) with full text and amendment tracking<\/p>\n
– Vector embeddings (Voyage AI, 1024d) for every case<\/p>\n
– Bilingual legal translation pairs across 11 Indian languages (Hindi, Tamil, Telugu, Bangla, Marathi, Gujarati, Kannada, Malayalam, Punjabi, Odia, Urdu) paired with English<\/p>\n
For context: India has the world’s largest common law system.<\/p>\n
40M+ pending cases. Court judgments are public domain under Indian law (no copyright on judicial decisions). But the raw data is scattered across 25+ different court websites, each with different formats, and many orders are scanned image PDFs with no searchable text.<\/p>\n
Available as:<\/p>\n
– REST API (sub-500ms hybrid semantic + keyword search)<\/p>\n
– Bulk export (JSON \/ Parquet)<\/p>\n
– Vector search via Qdrant<\/p>\n
The bilingual legal translation pairs might be interesting for NLP<\/p>\n
researchers working on low-resource Indian languages. Legal text is formal register with precise terminology, which is hard to find in most Indian language corpora.<\/p>\n
Details: vaquill ai <\/p>\n
Happy to answer questions about the data collection process, schema, or coverage gaps.<\/p>\n<\/div>\n
submitted by \/u\/zriyansh <\/a>
[link]<\/a><\/span> [comments]<\/a><\/span><\/p>
