Pith

Help

Sign in and copy the key from the account page. Opening that page registers the key on the extract server. Send it as Authorization: Bearer KEY. The key is 32 lowercase hex characters.

Extract

The body is the PDF itself. The largest file is 64 MB.

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

A success response is one JSON object per line. The headers are X-Document-Id, X-Pages, X-Deferred-Pages, and X-Scan-Pages.

{"page":0,"path":["Methods"],"text":"The sample was stored at room temperature.","left":72.0,"bottom":640.0,"right":480.0,"top":710.0,"table":false}

page starts at 0. path is the headings in force. The box is in PDF points, with the origin at the bottom-left. A ruled table sets table to true and joins cells with | .

A page with no text layer is {"page":4,"deferred":true}. When that page is a scan, OCR fills the line later and adds "ocr":true.

Account and documents

GET /v1/health needs no key. GET /v1/usage, with the key, returns pages used and the plan limit. At the limit, extract returns 429. Every page counts, including a deferred page. Delete does not return pages.

GET /v1/documents lists this key's files. GET /v1/documents/ID reads the same lines again. GET /v1/documents/ID/status is done, pending, or failed. DELETE /v1/documents/ID returns 204.

Start is 1,000 pages. Build is €19 a month for 25,000. Scale is €49 a month for 100,000.