▣ INGESTION KIT · DOCUMENT INTELLIGENCE KIT

Unified Document Extraction API

Hosted on ApyHub

What it does

Document Markdown Extractor turns an uploaded document into per-unit Markdown. Send a presigned_url, with optional filename and mime_type hints, and get back a document-order breakdown of units that you can join into the full text.

Each unit includes an index, kind, label, markdown, route, language, and warnings. That gives you both the extracted content and the context needed to rebuild page-by-page, slide-by-slide, sheet-by-sheet, or section-by-section output. The extractor also returns metadata with the detected MIME type and extraction model, plus diagnostics for OCR activity and embedded image handling.

Use Document Markdown Extractor when you need document content that is easier to search, chunk, review, or feed into downstream processing than the original file. It handles native text, OCR fallback, and embedded image text, while preserving stable warning codes such as ocr_fallback_used, table_split_into_blocks, and embedded_image_ocr_failed.

For PDFs, spreadsheets, slides, images, and text containers, you get a consistent Markdown representation without having to build your own parsing pipeline.

POST

Extract a document into per-unit Markdown

https://api.eu.apyhub.com/apyhub/unified-document-extraction

QUICKSTART

GUIDE

Quickstart

Send a presigned file URL so the API can extract its content.

curl -X POST "https://api.eu.apyhub.com/apyhub/unified-document-extraction" \
  -H "apy-token: $APY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "presigned_url": "https://assets.apyhub.com/samples/sample.pdf",
    "filename": "sample.pdf",
    "mime_type": "application/pdf"
  }'

What you'll get back

Returns a JSON object with a data object. data contains units (the extracted content units in document order), metadata (the detected MIME type and model), and diagnostics (OCR and embedded-image counts).

{
  "data": {
    "units": [
      {
        "index": 1,
        "kind": "page",
        "label": "Page 1",
        "markdown": "Adobe Acrobat PDF Files Adobe® Portable...",
        "route": "native",
        "language": "en",
        "warnings": []
      }
    ],
    "metadata": {
      "detected_mime_type": "application/pdf",
      "model": "apyhub-unified/v1"
    },
    "diagnostics": {
      "ocr_attempted": false,
      "embedded_images_total": 0,
      "embedded_images_ocrd": 0,
      "embedded_images_skipped_decorative": 0
    }
  }
}

TRY ITLIVE · 10 ATOMS

API key

Loading your default key…

The full key is used to call the gateway and stays in this tab — never sent to orbit or saved.

About this endpoint

What it does

Uploads a document via a presigned URL and returns its extracted content as per-unit Markdown, along with metadata and diagnostics about the extraction process.

Request Body

Parameter	Type	Mandatory	Description
filename	String	No	Optional. Used only as an extension hint for plain-text containers (`.csv`, `.tsv`, `.md`) and image naming.
mime_type	String	No	MIME type hint. Never trusted — routing uses the type detected from content.
presigned_url	String	Yes	S3 presigned GET URL (or any http/https URL). Fetched as-is; may exceed 2,048 characters. Never logged.

Response

Returns a JSON object with a data object field. data contains units, metadata, and diagnostics; units is an array of content units in document order, metadata describes the extraction engine and detected MIME type, and diagnostics reports OCR/image processing counters.

Parameter	Type	Mandatory	Description
data	Object	Yes	Container for the extraction result. Includes `units`, `metadata`, and `diagnostics`.
data.units	Object Array	Yes	The whole content contract, in document order. Join `markdown` in order for the full document. Each unit includes `index`, `kind`, `label`, `route`, `language`, `markdown`, and `warnings`.
data.units[].index	Integer	Yes	Contiguous 1-based position; maps to chunk `page_number`.
data.units[].kind	ENUM	Yes	Allowed values: `page`, `sheet`, `slide`, `image`, `section`.
data.units[].label	String	Yes	`Page N`, `Slide N`, sheet name, section heading, `Preamble`, `Document`, or CSV row range.
data.units[].route	ENUM	Yes	Allowed values: `native`, `ocr`, `mixed_ocr`, `native_plus_embedded_ocr`, `blank`. A unit with empty Markdown is always `blank`.
data.units[].language	String	Yes	ISO 639-1 code detected per unit; `und` when undetermined.
data.units[].markdown	String	Yes	Unit content. Tables are returned as Markdown tables; embedded image text appears as `> [Image text: …]` at its anchor.
data.units[].warnings	String Array	Yes	Stable snake_case codes such as `hidden_sheet`, `hidden_rows`, `hidden_slide`, `table_split_into_blocks`, `embedded_image_ocr_failed`, `ocr_fallback_used`, `ocr_failed_text_layer_used`, `ocr_attempted_no_content`, `ocr_not_attempted`.
data.metadata	Object	Yes	Extraction metadata. Includes `detected_mime_type` and `model`.
data.metadata.model	String	Yes	Extraction engine/model identifiers joined with `+`; output is deterministic per value.
data.metadata.detected_mime_type	String	Yes	Type detected from content; may differ from the `mime_type` hint.
data.diagnostics	Object	Yes	Always present diagnostic counters for OCR and embedded image handling.
data.diagnostics.ocr_attempted	Boolean	Yes	Indicates whether OCR was attempted.
data.diagnostics.embedded_images_total	Integer	Yes	Total embedded images processed or skipped.
data.diagnostics.embedded_images_ocrd	Integer	Yes	Count of embedded images OCR’d.
data.diagnostics.embedded_images_skipped_decorative	Integer	Yes	Count of embedded images skipped because they were decorative.

Notes

data.diagnostics is always present, and embedded_images_total follows the relationship embedded_images_total = embedded_images_ocrd + embedded_images_skipped_decorative + failed; any failed embedded-image OCR is also reported on the relevant unit with the embedded_image_ocr_failed warning.

▣ COMMON ERRORS

Errors any endpoint can return

400bad_request

Required parameter missing or malformed body.

401unauthorized

API key missing, revoked, or not authorized for this service.

429rate_limited

Your plan's per-second rate exceeded. Retry with exponential backoff.

503upstream_busy

Backend temporarily unavailable. Try again in a few seconds.

Similar APIs

Closest matches in the catalog based on this service.

CALLImage & PDF OCR to Markdown APICallable Labs CALLExtract Text from PDF Document APIDosvak LLC CALLAI Document Data Extraction APIApyHub CALLAI-ready Clean Data Extractor APIApyHub

From the blog

Reads that touch this API.

PDF to Markdown with Python: A Resumable OCR Pipeline for RAGOct 25, 2026 Understanding Markdown and How to Convert It: A Developer’s GuideJul 9, 2026 Top 5 Use Cases: Extract Text from PDF for your applications and workflowsJul 9, 2026 Top Tools, APIs and Use Cases for Extracting Text from Any WebsiteAug 10, 2026