run-llama/liteparse
Summary
LiteParse is an OSS PDF parsing tool from run-llama, focused on fast, local parsing with high-quality spatial text data and bounding boxes. It supports multiple languages and bindings (Rust core, Node, Python, WASM) and offers OCR via bundled Tesseract or HTTP OCR servers, plus a multi-format input flow and screenshot generation for LLMAgents.