liteparse

Local document and PDF parser (Rust core, Python/Node bindings) that returns spatial text with bounding boxes. It supports PDFs, Office files, and images, runs OCR via bundled Tesseract, and renders page screenshots. The skill outputs layout-preserved plain text or structured JSON with per-page text items and bounding boxes.

Tasks

Parse a scanned PDF to structured JSON with bounding boxes using the lit CLI
Batch-parse a directory of PDFs to JSON with OCR disabled using lit batch-parse

Inputs

.pdf files
.docx, .xlsx, .pptx, .doc, .odt files
.png, .jpg, .tiff, .webp, .svg files
PDF bytes or stdin

Outputs

Layout-preserved plain text
Structured JSON with per-page text_items, bounding boxes, font metadata, and confidence scores
PNG page screenshots
Combined text items and bounding boxes for phrase searches

Limitations and checks

Does not produce Markdown
Does not call cloud LLMs
Office and image files require LibreOffice and ImageMagick respectively, and fail if missing
Office file bytes cannot be parsed directly; they require a file path and conversion
Verify your installation by running lit --help and importing liteparse in Python to print the version
Confirm LibreOffice (soffice) is on PATH before parsing Office files
Confirm ImageMagick (convert or magick) is working before parsing image files
Test OCR quality on a sample scanned page and adjust DPI or ocr-language if results are poor

AI Search

Find projects, verify facts, compare options, or turn a complex need into an actionable plan

Try a searchA click only fills the search box; you stay in control

Collection Scenario Project type License Deployment Resource form

liteparse

Tasks

Inputs

Outputs

Limitations and checks

AI Search

Project Details