Coming soon
There is no API reference to publish yet
A reference describes something you can call. Nothing here can be called, so there is nothing to document.
The OCR API page describes what is being designed. It accepts no requests.
The API is not operational. No endpoint accepts requests, no key is issued, and API access is not included in any active plan. This page is not a beta gate or a waiting list.
Why this page is almost empty
It used to carry a base URL, an authentication header, five endpoints, request and response examples and rate-limit headers. All of it described a service that does not exist, and a reader could reasonably have begun an integration against it.
Publishing a plausible reference for something unbuilt is worse than publishing nothing: it costs the reader real work before they discover the gap. The examples were removed rather than qualified with a warning.
When the service exists and has been tested, a real reference will replace this page. Until then there is no date, because a date given before the work is done is a guess.
What works today
Extraction runs in the browser. Invoice OCR, the PDF invoice parser, Receipt to Excel and Image to Excel all work now and export Excel, CSV or JSON — as a user-facing workflow rather than a service you can call from your own code.