Search Guide

This tool lets you search across the publicly available court filings from Indiana v. Richard M. Allen using natural language queries. Instead of opening documents one at a time, you can ask a question and the system finds the most relevant passages across every indexed filing.

The indexed documents include motions, orders, transcripts, hearing records, exhibits, and other filings from the public docket. Documents that are sealed, redacted, or image-only (scanned without OCR) are not included in the search results.

This is a semantic search engine, not a keyword search. It understands meaning, not just exact words. When you type a query, the system converts it into a mathematical representation and compares it against every document chunk in the database, returning the closest matches by meaning.

This means you can search for concepts, not just terms. A query like "defense arguments about the bullet" will find relevant passages even if those passages don't contain the exact word "bullet"; they might say "round," "cartridge," "projectile," or "firearms evidence."

Each result shows the source filing, date, a relevance score, and a link to the original document on Google Drive.

Natural language works best. Ask questions or describe what you're looking for in plain English.

General guidance:

• Use full phrases and questions rather than single keywords
• Include context: "search warrant for Whiteman Drive" works better than just "search warrant"
• The relevance percentage is a rough guide. Anything above 50% is usually on point; below 40% may be tangential
• Adjust the results count (3, 5, 10, or 15) to see more or fewer matches
• Click "Expand" to see the full text of a result, or "View Source" to open the original filing

The search covers text extracted from 667 PDF filings downloaded from the public Google Drive case file archive. This includes:

• Court orders and rulings
• Defense and prosecution motions
• Hearing and trial transcripts
• Affidavits and search warrants
• Appellate filings
• Exhibits (text-based only; images and scanned pages without embedded text are not searchable)

Approximately 300 files in the collection are image-only PDFs (scanned documents, photographs, handwritten letters) that couldn't be text-extracted. These are still available through the case file archive but won't appear in search results.

• This is a search tool, not legal analysis. Results are document excerpts, not conclusions.
• Sealed and confidential filings are not included.
• Some document text may have extraction artifacts (spacing issues, broken words) from PDF conversion.
• The search API may take a few seconds to respond on the first query if the server has been idle.
• Results show document chunks (roughly paragraph-length passages), not full documents. Use the "View Source" link to read the complete filing.

This tool was built as a proof of concept for document intelligence applied to public court records. The underlying technology (semantic vector search over extracted document text) is designed to be adaptable to any large collection of court filings.

All documents are sourced from publicly available court records. This project is not affiliated with the court, any party to the case, or any legal organization.