Extract structured Markdown, JSON, and HTML data from PDF documents.
npx skills add https://github.com/chujianyun/skills --skill opendataloader-pdf
Please help me install this Agent Skill. Skill: opendataloader-pdf Source: https://github.com/chujianyun/skills/tree/main/skills/media/opendataloader-pdf Command: npx skills add https://github.com/chujianyun/skills --skill opendataloader-pdf
Extracting structured data from PDFs for RAG pipelines or LLM training is painful: tables break, reading order gets scrambled, and scanned pages return nothing. This Skill wraps OpenDataLoader PDF, a benchmark-leading parser, to convert PDFs into clean Markdown, JSON with bounding boxes, or HTML.
Install opendataloader-pdf with pip and ask the agent to convert a PDF file into Markdown and JSON output in a chosen directory.
High-intent search queries and answers about installing and using this skill.
Install opendataloader-pdf with pip, then call opendataloader_pdf.convert with the input path, output directory, and format set to markdown. The CLI equivalent is opendataloader-pdf input.pdf output_dir/.
OpenDataLoader PDF scores 0.90 overall on the opendataloader-bench benchmark, ahead of docling, marker, and mineru. It outputs Markdown and JSON with bounding boxes, which suits chunking and source citation in RAG pipelines.
Yes, scanned documents are handled in hybrid mode by adding --force-ocr to the command. For non-English scans, pass language codes with --ocr-lang, for example "ko,en".
Use the JSON output format, where each element includes a bbox array with x0, y0, x1, y1 coordinates plus its type and content. Tables are returned as HTML inside the JSON structure.
Each convert() call starts a new JVM process, so looping over files repeatedly pays that startup cost. Pass all files or folders in a single convert call to amortize JVM startup.
Use the default fast mode for standard digital PDFs. Switch to hybrid mode with docling-fast for complex borderless tables, scanned pages, or documents containing math formulas.
googleworkspaceTransfer attachments from email to specified Google Drive folders.
langchain-aiOrchestrate document loading, splitting, embedding, and vector store management for RAG pipelines.
skaile-aiExtract structured knowledge from scientific papers and generate research folders.
openaeonExtract text, Markdown, and tables from PDFs and scanned documents.
