liteparse_samples: Interactive Parser Demos from LlamaIndex

liteparse_samples is a set of interactive demos for LiteParse — a fast, local, model-free document parser from LlamaIndex. The repo includes three standalone demos built around real government and financial PDFs.

What’s included

• Parser comparison — Side-by-side output from LiteParse, PyPDF, and PyMuPDF across 8 document sections from 5 real-world PDFs (FDIC, Federal Reserve, CMS, IRS, WHO). Includes per-document timing for each parser.
• Visual citations — Exact keyword search over parsed documents with bounding boxes highlighted directly on the source PDF page image.
• Research docs — A Claude Code skill (/research-docs) that parses documents, answers questions with AI, and generates an HTML report with cited source pages highlighted.

Each demo is configured via a docs.json file — add your own PDFs and pages to customize. Requires Python 3.9+.

Repo: github.com/jerryjliu/liteparse_samples · Docs: developers.llamaindex.ai/liteparse