PDF Extractor API: Text, Markdown, Tables, No Login
Pricing
from $1.20 / 1,000 pdf extractions
Extract text from public PDF URLs: plain text, Markdown, tables, links, outline, page count and metadata per PDF or per page, with scanned pages flagged. $2 per 1,000 PDFs ($1.60 on Starter). Failed files are free. No login.
Add the public web addresses of PDF files, one per line. Each PDF becomes one dataset row (or one row per page, see Row per).
Pick an Apify Dataset that holds your PDF addresses, one entry per item. They are added to the entries typed above.
Name of the dataset field that holds the entry. Leave empty to use the first text field it finds.
Web address of a .txt, .csv or .json file with your PDF addresses, one entry per line. A header row is skipped. Up to 10 MB.
Choose one row per PDF (full text, tables, links and metadata together) or one row per page (text and tables of each page, for page citations and chunking).
Return the plain text in reading order, with line breaks as in the PDF and a form feed between pages.
Return Markdown with headings, lists and tables, with running headers, footers and page numbers removed. Ready for LLMs and RAG pipelines.
Detect tables and return each one as rows of cells, in addition to the Markdown table.
Also save each PDF's text and Markdown as files in the key-value store, linked in textFileUrl and markdownFileUrl (with Row per: Document). Very long documents are always saved this way.
Read at most this many PDFs from the list.
Read at most this many pages of each PDF, starting from page 1. The row says truncated: true when a PDF has more.
Download PDFs up to this size. Larger files return status too_large and are free. For files over 50 MB, give the run 2 GB of memory or more.
Read up to this many PDFs at the same time (at most 2 per website). Lowered automatically when large files would not fit in the run's memory.
