liteparse
Local document and PDF parsing that returns spatial text with bounding boxes
by k-dense-aiFree
Description
liteparse is an agent skill for local document and PDF parsing that returns layout-aware text with bounding boxes. It is for extracting text from PDFs, DOCX and other Office files, and images, and for running OCR on scanned pages. LiteParse is an open-source parser with a Rust core and Python and Node bindings. It can produce layout text, structured JSON, or heuristic Markdown. Its spatial text items may span several words, and the Python option emit_word_boxes adds word boxes when you need them. The verified release is the Python liteparse 2.15.0 from 29 September 2026, with the command-line tool and synthetic PDF and image test files checked on Python 3.13. The Node and Rust examples were checked against source but not run. Images are converted with bundled Rust libraries, not ImageMagick. The skill says no cloud account is needed for parsing. It also notes two data flows to be aware of. Missing Tesseract language data can be downloaded from GitHub, and an HTTP OCR service receives document images only if you explicitly configure one. The skill includes a reference file for choosing between this parser and other options such as MarkItDown and LlamaParse.…
Highlights
- Layout-aware text with bounding boxes
- PDF, DOCX, Office and image input
- OCR for scanned pages
- Layout text, JSON or Markdown output
- Runs locally, no cloud account
Ratings & Reviews
No ratings yet
Be the first to rate liteparse.
People also visit
liteparse FAQ
What does it extract?
Text from PDFs, DOCX and other Office files, and images, with layout and bounding boxes. It can run OCR on scanned pages.
What output formats?
Layout text, structured JSON, or heuristic Markdown.
Does it send my documents anywhere?
Not by default. The skill says no cloud account is needed, but an HTTP OCR service receives document images only if you configure one, and missing Tesseract language data can download from GitHub.
Which version was verified?
The Python liteparse 2.15.0 release from 29 September 2026, tested on Python 3.13.
How do I install it?
Run npx skills add https://github.com/k-dense-ai/scientific-agent-skills --skill liteparse.
Is it checked for security?
Its security audits show passes from Gen Agent Trust Hub, Socket and Snyk.
What are the best liteparse alternatives?
Popular alternatives to liteparse include markitdown, data-scraper-agent and django-celery. See all liteparse alternatives