Parse your first document

Transform turns documents into parsed content and structured data. Use the web app walkthrough below, or skip to your first API request to start with code. You do not need to complete both paths.

Upload a document

Upload your own document or download the ACME Corp Annual Report sample PDF to try the flow. Check the supported file types if you are unsure whether your file is supported.

Parse the document

Choose Parse. When processing finishes, review the parsed content alongside your document.

Extract structured data (optional)

Choose Extract if you need specific fields as structured data. Define the fields you need, then review the extracted values.

Use Get Code

Open Get Code to carry the flow into an API integration. Use your API key when running the code in your application. See integrate with Get Code for the next steps.

Review the output

Compare the parsed content and any extracted values with your source document. See result interpretation for details. See recovery guidance if an upload or processing step fails.

Make your first API request

You can Parse a document directly through the API without uploading it in the web app first. The -i option prints response headers so you can retain Location if processing continues. Use your API key in UNSTRUCTURED_API_KEY and a supported document saved as document.pdf. No workspace parameter is required. To try the example without your own file, download the ACME Corp Annual Report sample PDF and save it as document.pdf in your project directory.

export UNSTRUCTURED_API_KEY="your-api-key"

curl -i https://transform.unstructured.io/api/v2/parse \
  -H "unstructured-api-key: $UNSTRUCTURED_API_KEY" \
  -F "input=@document.pdf"

A synchronous request keeps the connection open until parsing finishes. Multi-page documents can take about a minute or longer. That wait is expected. For an immediate response, see asynchronous processing. HTTP 200 returns a Parse result directly. This response was captured from the ACME sample PDF on September 12, 2026. The Markdown is shortened and the result ID is replaced with an example UUID:

{
  "id": "11111111-1111-4111-8111-111111111111",
  "status": "completed",
  "profile": "balanced",
  "markdown": "## ACME CORP\n\n## Annual Financial Report \u2014 Fiscal Year 2025\n\n\u2026",
  "format_version": "2.0",
  "metadata": {
    "page_count": 10
  },
  "extracted_data": [],
  "warnings": [],
  "elements": [],
  "source": {
    "file_id": "file_example",
    "filename": "document.pdf",
    "mimetype": "application/pdf",
    "expires_at": "2026-09-13T00:00:00Z"
  }
}

Read the parsed text in markdown. Save id to Extract from the same Parse later. Check status and warnings before using the output. metadata describes the document, format_version identifies the result format, and source identifies the input file and its expiry. The source values in this shortened example are representative. This Markdown example has empty elements and extracted_data arrays. See Parse response format for other output options.

​Review the output

​Make your first API request

Review the output

Make your first API request