Pith

Born-digital PDFs, back as citation-ready chunks.

Pith extracts text from a born-digital PDF. Each chunk carries the page, the heading, and the box, so a citation can point at the words. A 196-page digital PDF took about one second of server time on one core.

Call

curl -D - --data-binary @file.pdf \
  -H "Authorization: Bearer KEY" \
  https://pith.grbsystem.com/v1/extract

What comes back

The response is one JSON object per line. How a born-digital PDF is extracted.

  1. 1

    The page

    page starts at 0, so a citation can name the page the words came from.

  2. 2

    The heading

    path is the headings in force for that chunk, outer heading first.

  3. 3

    The box

    left, bottom, right, and top are PDF points from the bottom-left of the page.

Plans

The key stops when the plan's pages are used. Cancellation is from the account page.

Build

€19/ month

  • 25,000 pages
  • The same API
  • Paid with Stripe
Choose Build

Scale

€49/ month

  • 100,000 pages
  • The same API
  • Paid with Stripe
Choose Scale