NEW AGENT SKILL

Stop wasting your context window on PDF extraction

Extracting Markdown from PDFs takes precision. Nutrient’s new pdf-to-markdown skill gives Claude and Codex fast, accurate extraction, so your AI workflow starts with clean Markdown instead of noisy output that eats into your context window.

GitHub repo
INSTALL

Install for Claude Code and Codex, or use the CLI

Install Nutrient’s new PDF-to-Markdown skill in Claude Code with the Nutrient marketplace, use it in Codex, or install the standalone CLI.

CLAUDE CODE

Install it in Claude Code

We published pdf-to-markdown in the Nutrient marketplace so you can install it directly in Claude Code. 

/plugin marketplace add pspdfkit-labs/nutrient-skills
/plugin install pdf-to-markdown@nutrient-skills
CLI

Install the standalone CLI

Or install the standalone pdf-to-markdown CLI tool directly. 

curl https://agent-cdn.nutrient.io/pdf-to-markdown/install.sh | sh
BENCHMARKS

Speed and accuracy shouldn’t be a tradeoff

Better PDF-to-Markdown extraction means less cleanup, fewer wasted tokens, and more reliable downstream results for Claude and Codex.

Extraction accuracy
Score per solution
0.88 docling
0.88 Nutrient
0.83 opendataloader
0.82 pymupdf4llm
0.59 markitdown
0.58 pypdf
0.57 liteparse
Extraction time per page
In seconds per solution
0.007 Nutrient
0.014 opendataloader
0.019 pypdf
0.106 markitdown
0.233 liteparse
0.252 pymupdf4llm
0.618 docling
Reading order (NID)
Score per solution
0.92 Nutrient
0.90 opendataloader
0.90 docling
0.88 pymupdf4llm
0.87 pypdf
0.86 liteparse
0.84 markitdown
Table structure (TEDS)
Score per solution
0.89 docling
0.66 Nutrient
0.49 opendataloader
0.48 pymupdf4llm
0.27 markitdown
0.00 liteparse
0.00 pypdf
Heading level (MHS)
Score per solution
0.82 docling
0.81 Nutrient
0.77 pymupdf4llm
0.74 opendataloader
0.00 markitdown
0.00 liteparse
0.00 pypdf
Faster with Nutrient
How much faster Nutrient is
91× FASTER THAN docling
37× FASTER THAN pymupdf4llm
34× FASTER THAN liteparse
16× FASTER THAN markitdown
3× FASTER THAN pypdf
2× FASTER THAN opendataloader

Benchmark data recorded on Apple M3 Ultra (no GPU)

TRUST

Free to use, and your PDFs don’t go to Nutrient

Use the extractor freely within the included usage limits, and keep your documents inside your own workflow.

USAGE

Free for up to 1,000 documents per month

Free for processing up to 1,000 documents per calendar month.

Commercial license required for: processing more than 1,000 documents per month, redistributing the binary, or OEM/white label use.

Contact sales
PRIVACY

Your PDFs stay in your workflow

The skill reads your PDFs and produces Markdown output without uploading your documents to Nutrient.

  • Local PDFs aren’t routed through a Nutrient service
  • Claude, Codex, and any other model provider still apply their own policies