Upstage Document Parse icon

Upstage Document Parse

Convert documents into structured HTML/Markdown using Upstage Document Parse

Actions4

Sync Parse (Upload File)

AI-generated

Summary

Parse an uploaded document via Upstage's document-digitization API, returning structured content per selected return mode and output formats.

Inputs

  • binaryPropertyName (required) — Name of the input item binary property that contains the file. Defaults to 'data' if not specified.

Output shape

Returns an array of paired items. Depends on returnMode: 'full' returns the complete API response JSON; 'content_html' returns { html: string }; 'content_markdown' returns { markdown: string }; 'content_text' returns { text: string }; 'elements' returns { elements: array }. For asyncSubmit, returns { request_id: string, submitted: boolean }.

returnMode display options are only shown for the sync operation.

Examples

Example 1: Parse a PDF and get HTML layout

binaryPropertyName: data; model: document-parse; outputFormats: [html]; returnMode: full

Example 2: Extract text content from an image with OCR

binaryPropertyName: data; model: document-parse; ocr: force; outputFormats: [text]; returnMode: content_text

Example 3: Generate base64-encoded cropped images for tables and figures

binaryPropertyName: data; model: document-parse; base64Categories: [figure, table]; outputFormats: [html, markdown]; returnMode: full

Links

Discussion