Skip to content

vision-parse

Parse PDFs into Markdown with vision language models

Visit official site

What vision-parse is

The tool is described as a PDF parser that uses vision language models to produce well-formatted Markdown. It recognises text and tables while keeping the document structure.

What you can do with it

  • Extract text and tables from a PDF
  • Keep hierarchy, style and indentation
  • Process multi-page PDF documents

Getting started

  • Install Python 3.9 or newer
  • Install the package with pip
  • Add optional model dependencies
  • Create a VisionParser instance

Pros & cons

✓ What we like

  • Local model hosting supports offline and secure processing
  • Parameters can be tuned for detailed extraction

! What to watch out for

  • Python 3.9 or newer is required
  • Optional OpenAI or Gemini dependencies may need installing

FAQ

Which models are supported?

OpenAI, Google Gemini and Ollama are listed.

Can it run locally?

Ollama supports local and offline document processing.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review