Skip to main content

Document Parsing

Real-world data is trapped in PDFs, Word docs, and messy HTML.

Unstructured.io​

A unified API for parsing PDFs, PPTX, HTML, etc., into clean JSON.

LlamaParse​

Specifically designed to parse complex PDFs with tables and charts into Markdown for LLM ingestion.

Surya OCR​

An open-source, highly accurate multi-lingual OCR model that outperforms Tesseract.

HTML to Markdown​

Always convert HTML to Markdown before feeding it to an LLM. It saves tokens and removes noise.