Multilingual Indian Legal Corpus for Supervised Fine-Tuning

A large-scale multilingual Indian legal dataset curated by Prarabdha, last updated in October 2025. It contains text from publicly available legal sources across multiple Indian languages. The corpus is structured to support research in domain-specific language model training and legal workflow automation.

Use Cases

• Supervised fine-tuning of LLMs for legal text generation using the multilingual corpus.
• Training models for legal question answering based on the structured text data.
• Developing case retrieval systems by leveraging the document-level legal text.
• Building policy reasoning agents using the curated legal domain text.
• Creating agentic systems for legal workflow automation from the processed corpus.

Strengths

• Multilingual corpus covering at least 9 Indian languages including English, Hindi, and Tamil.
• Dataset size is categorized as 1M to 10M entries based on platform tags.

Limitations

• Specific row count, column structure, and sample data are unavailable for assessment.
• Potential geographic bias as data is sourced exclusively from Indian legal sources.

Provenance

Source

Publicly available legal sources across India.

Collection Method

Curated and processed from public legal texts.

Time Range

null

Freshness

Last updated on 2025-10-29.

Geography

India