Claude quickstart - Unstructured

We’ll use an example company annual report to demonstrate Unstructured’s parsing capabilities.

What's in the example file?

The file we’re going to parse contains a range of unstructured text and graphic elements, such as tables, charts, graphs, and handwriting.

Parse the file

Estimated time from login to opening the parse results file: about 5 minutes

Open Claude

Connect Claude to the Transform MCP

Click here to open the Unstructured Transform connector page, and then click Connect.

Can't find the connector?

If you’re a Team or Enterprise account user, your plan Owner needs to add Unstructured Transform to your organization before you can choose it. Have them do this:

Click here to open your Organization settings > Connectors page and then click Browse connectors.

Search for the Unstructured Transform connector, select it, and click Add to your team.

Create your Unstructured account

When the Unstructured browser window opens, click Register and follow the on-screen directions.

Confirm the connection

After you’ve finished, click here to open the Connectors dialog and verify that Unstructured Transform appears in the list.Close the Connectors dialog.

Download the file

Download the ACME Corp Annual Report.

Parse the file

Give Claude the following prompt:

    Use the Unstructured Transform MCP server to parse a file, 
    and return the results as JSON. Use the open_file_manager tool 
    to enable me to select the file. Give me instructions for 
    how to use the interactive card. When the JSON is ready, 
    open the downloader card so I can grab it.

Follow the instructions in Claude to upload the file and have Unstructured parse it.Download the JSON results once the job finishes.

Review the results

Unstructured generates a JSON file that is a collection of elements it found in the file: text, tables, images, titles, headers, footers, and more. Each element has a unique ID, as well as a a field naming the file it came from.

The JSON results file is minified by default.Most browsers contain a pretty print option that displays the JSON in human-readable form with proper indentation and line breaks. For example, in Google Chrome, open the file and then check Pretty print; in Mozilla Firefox, open the file, click Raw Data, and then Pretty Print.

Some common document element types

Here’s a partial list of the document element types you’ll see in your parsed JSON results:

Element type	Description
`Footer`	Captures document footers.
`Header`	Captures document headers.
`Image`	A text element for capturing image metadata.
`ListItem`	A `NarrativeText` element that is part of a list.
`NarrativeText`	An element consisting of multiple, well-formulated sentences. This excludes elements such as titles, headers, footers, and captions.
`PageBreak`	Captures page breaks.
`PageNumber`	Captures page numbers.
`Table`	An element for capturing tables.
`Title`	A text element for capturing titles.
`UncategorizedText`	Base element for capturing free text from within files. Applies to extracted text not associated with bounding boxes if the input is a PDF file.

Unstructured also includes metadata fields inside elements so that the parsed JSON represents a rich and accurate capture of your file. For example:

The metadata for each element includes coordinates for its position on the page.
If an element resides in another element, Unstructured includes a parent_id in the child element to retain this relationship.
For Table elements, Unstructured includes escaped JSON that represents a complete rendering of the table in HTML. To find a table element within your file, search for the string text_as_html. To visually render Unstructured’s HTML table representations, you can render the resulting HTML (in an online tool such as Div Table or HTML-Online).
For Image elements, Unstructured includes a Base64 representation of the image. To find this within your file, search for the string image_base64. To convert the Base64 representation back to the original image, paste the contents of an image_base64 element into an online tool such as Base64 Guru.
The above links to third-party websites are provided solely as a convenience. We do not control, approve, or endorse the content, products, or services offered on these external sites. We assume no responsibility for your use of these external sites.

Go further: Extract structured data

Estimated time from downloading the file to opening the extracted data file: about 5 minutes Now let’s see how Unstructured lets you control exactly what information gets extracted from your files, and the structure, or schema, to apply to that data. You can specify the data structure you want, or have Unstructured recommend a schema based on the information it finds in your files. We’re going to have Unstructured parse one example form and suggests a schema for the data it finds.

What's in the example file?

This medical questionnaire mixes tabular data, free-form text, and handwriting in several fonts and colors.

Download the file

Download the example medical form.

Extract data from the file

Give Claude the following prompt:

    Use the Unstructured Transform MCP server to extract 
    structured data from a file. I don't have a schema, 
    so suggest one based on what's in the document, show it to me, 
    and then run the extraction with it. Use the open_file_manager tool 
    to enable me to select the file. Give me instructions for 
    how to use the interactive card. Give me the extracted data 
    as a JSON file I can download.

Follow the instructions in Claude to upload the file and have Unstructured parse it.Accept the schema Unstructured suggests, and download the JSON results once the job finishes. (By default, Unstructured returns the extracted JSON inline, but the above prompt asks for a downloadable file to make reviewing the results easier.)

Review the structured data

When we parsed the company report sample file, Unstructured generated elements based on how the file presents the data. That is, a title, a header, a table, and so on. With this medical form, Unstructured extracted the meaning of the data into a format you defined: full name, date of birth, home address, and so on. Unstructured reads the handwritten, scanned form and returns a concise data structure. It groups and nests the fields according to your specifications, so your systems can consume the output directly. While this was a single form, imagine if you had a stack of these patient forms in PDF format. This is the kind of job Unstructured is built to process at scale.

Next steps

Use other AI tools

Use Unstructured with Claude Desktop, Claude Code, Cursor, Codex, Visual Studio Code, and more.

More output options

Use plain language to tweak Unstructured’s output format.

Prompting tips and tricks

Suggestions and best practices for getting Unstructured to do what you want.

Advanced extraction techniques

Extend this quickstart to extract custom data from your files in additional ways.

Auto-generate production-ready code

Ask Unstructured for the Python or cURL script that produces this output, then add it to your development environment.

​Parse the file

​Review the results

​Go further: Extract structured data

​Review the structured data

​Next steps

Use other AI tools

More output options

Prompting tips and tricks

Advanced extraction techniques

Auto-generate production-ready code

Parse the file

Review the results

Go further: Extract structured data

Review the structured data

Next steps