# docfold

> Unified Python interface over 20+ document-structuring engines — PDF, DOCX,
> XLSX, PPTX, HTML, and images → structured, LLM-ready markdown/JSON, with
> extension-aware routing and automatic engine fallback. Open source, MIT.
> Ships a CLI (`docfold`) and an MCP server (`docfold-mcp`).

## Setup

- [Agent setup instructions](https://mihailorama.github.io/docfold/install.md):
  install, verify the CLI, and register the MCP server in one command
  (`docfold install claude|codex|cursor|vscode|generic`).

## Docs

- [README](https://github.com/mihailorama/docfold/blob/main/README.md)
- [Evaluation framework](https://github.com/mihailorama/docfold/blob/main/docs/evaluation.md)
- [Benchmarks](https://github.com/mihailorama/docfold/blob/main/docs/benchmarks.md)
