MerchantryTidbits

doc-parsing / library

unstructured

Capability: unstructured

Use it when

  • partition PDF HTML into structured elements
  • chunk documents for RAG
  • parse messy docs before embed
  • document structure for retrieval

What it solves

pdf-table-garbledocr-neededembedding-recall-poor

Not the fit when

  • bot-detection-challenge

Install

pip install unstructured

Invoke

partition_pdf / partition_html to structured elements before chunk+embed.

Alternatives

No reviewed alternatives recorded yet.