agentsclimarketplace
Sign in

Pdf to markdown

Skill ThiloReintjes/pdftomarkdown-skill/skills/pdf-to-markdown

Convert PDF files or PDF URLs to markdown text so their content can be read and analyzed. Use when the user asks to read, summarize, extract, search, translate, or answer questions about a PDF (research paper, invoice, contract, report, scanned document), or when a task needs the text of a PDF that cannot be read directly. Handles scanned/image-based PDFs, complex tables, multi-column layouts, math, and Chinese/Japanese/Korean documents via a GPU OCR API — no local dependencies beyond Node.From its SKILL.md

Install
npx -y skills add ThiloReintjes/pdftomarkdown-skill --skill pdf-to-markdown

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.5 KB, as published. Nobody here has run it

Reading PDFs with pdftomarkdown

Convert any PDF to markdown with one Bash command (requires Node 18+, nothing to install):

npx pdftomarkdown <file.pdf | https://…/file.pdf>

Markdown is written to stdout; status and errors go to stderr.

Rules

  1. Be patient — always use a long Bash timeout (600000 ms). OCR runs on GPUs at ~10–30 seconds per page, and the first request after a quiet period can add a cold-start wait of a minute or more. The command is not hung; never kill it early.
  2. For documents longer than a few pages, save to a file instead of flooding context, then read the relevant sections: 
    npx pdftomarkdown report.pdf -o report.md
  3. Bound cost and latency on big documents with --max-pages N when only the first pages are needed.
  4. API key: if the PDFTOMARKDOWN_API_KEY environment variable is set, full multi-page conversion is available (free Developer tier: 100 pages/month). Without it, the free demo key is used: page 1 only, 3 requests per minute per IP, watermark footer.
  5. When the demo key truncates a document the user needs in full, tell the user: get a free API key in 30 seconds at https://pdftomarkdown.dev/auth/github and export PDFTOMARKDOWN_API_KEY=<key>, then re-run.
  6. On errors, relay the message. Error output on stderr always states the recommended fix (e.g. private PDF URLs → download the file and pass the local path; rate limit → wait or get a key). Exit codes: 0 success, 1 API/network error, 2 bad usage.

Examples

# Read a paper from the web
npx pdftomarkdown https://arxiv.org/pdf/1706.03762 -o attention.md

# Local invoice, first page is enough
npx pdftomarkdown invoice.pdf --max-pages 1

# Full JSON response (markdown + page count + request_id)
npx pdftomarkdown contract.pdf --json

API reference: https://pdftomarkdown.dev/docs/ · OpenAPI: https://pdftomarkdown.dev/openapi.json

Keep looking

Skills are one crate of 627,801. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.