---
title: OCR Document Data Extraction API
slug: extract-read-data
url: https://apyhub.com/apyhub/service/extract-read-data
provider: ApyHub
categories: [Artificial Intelligence, Data Extraction]
tags: [document-extraction, ocr, pdf-extraction, layout-analysis, azure-document-intelligence]
auth: api_key
---

# OCR Document Data Extraction API

Extract reading data from a file upload or URL. Returns ApyHub or Azure parsing output for document workflows and indexing.

## Endpoints

| Method | Path | Description | Atoms |
| --- | --- | --- | --- |
| POST | `/ai/document/extract/read/file` | What it does Uploads a file and returns extracted read-data parsing output. You can choose whether… | 500 |
| POST | `/ai/document/extract/read/url` | What it does Extracts read data from a document at the provided URL and returns parsing output in a… | 500 |

## About

## What it does
Read Data Extractor pulls text-layout parsing results from a PDF or other readable file, or from a file URL. Send a binary file to `POST /file` or a `url` to `POST /url`, and choose whether you want the `apyhub` result or the `azure` result with `requested_service`.

The service accepts optional Azure credentials when you want to run against Azure Document Intelligence. For file uploads, you can pass `azure_key`, `azure_region`, `azure_endpoint`, `azure_account_id`, and `azure_access_token`. For URL-based extraction, the `azure` object can include `key`, `region`, `endpoint`, `account_id`, and `access_token`.

The response is returned under `data`, with separate `azure` and `apyhub` objects. `azure` contains the raw Azure Document Intelligence parsing output, while `apyhub` contains the structured ApyHub Read Layout Transcription parsing output. That makes it useful when you need to compare provider output, route documents through an internal extraction pipeline, or feed parsed document content into downstream review and indexing workflows.

Use Read Data Extractor when you need programmatic access to extracted document reading data without building your own ingestion and parsing layer.

## Usage

Authenticate with an ApyHub API key in the `apy-token` header.
Full docs and a live playground: https://apyhub.com/apyhub/service/extract-read-data