Read

Read a photo of a filled paper copy → its values (3 credits a page)

Each field is cut out of the photo and read on its own; every reading goes through the same checks as a fill. Only pages that line up with the form are charged. Test keys: up to 10 pages a day. Does not make a PDF — send the values to /engine/fill.

post/api/v1/engine/scan/{id}

Request

  • Base URL: https://philforms.com
  • URL: https://philforms.com/api/v1/engine/scan/{id}
  • Auth: one of:
    • HTTP bearer
    • API key in header X-PDFPhile-Embed

Path parameters

idstring required

Query parameters

crops'1'

Return each field crop (data URI) for review — paid plans only

reader'local' | 'hybrid' | 'ai'

local = no AI (OCR + a digit network, decoded with each field's format); hybrid = local first, a model only for doubtful fields (default when a model is configured); ai = a model reads every field

Request body

imagestring

page 1 only

Response

{ data, confidence, rejected, unread, skipped, registration, quality } — confidence ≥ 0.9 = sure; quality[].level "retake" says why a clearer photo would read better. Review, then send data to /engine/fill

Changes

No recorded changes to this operation across all 1 revision of this API.