Thread by @AndyMarlowg

Spent more time cleaning PDF output than building the actual AI workflow.

So I tested @nutrientdocs PDF-to-Markdown CLI.

Here's what happened ↓

If you're building with RAG, AI agents, or knowledge bases, you've probably hit the same problem:

PDFs are full of messy layouts, broken tables, and formatting issues.

Before feeding documents into an LLM, you usually spend time cleaning everything manually.

I tried @nutrientdocs PDF-to-Markdown CLI to see how well it handles the conversion process.

The tool takes a PDF and converts it into structured Markdown that's much easier for:

• RAG pipelines
• LLM ingestion
• Documentation systems
• AI workflows

No complicated setup required.

What stood out was the reduction in cleanup work.

Instead of spending time fixing formatting issues after extraction, the Markdown output was already organized enough to drop into my workflow with minimal editing.

That's especially useful when processing large document collections.

If you work with PDFs and AI tools, this is worth testing yourself.

Check out the open-source repo:


Curious to see how it performs on different document types and real-world datasets. 🚀github.com/PSPDFKit/pdf-t…