LiteParse
Local PDF and office document parser with OCR and bounding boxes
What it is
LiteParse extracts text from PDFs, Word, Excel, PowerPoint files and images entirely on your machine, using PDFium for spatial text layout and Tesseract or any HTTP OCR server for scanned pages. It outputs Markdown, JSON with bounding boxes, or layout-preserving plain text, and can render page screenshots and flag documents that need heavier processing. Developers building RAG and LLM ingestion pipelines call it from Rust, Python, Node.js, the browser via WASM, or the command line.