> ## Documentation Index
> Fetch the complete documentation index at: https://learn.workflow.dog/llms.txt
> Use this file to discover all available pages before exploring further.

# Extract Text

> Pull a configured set of text fields from unstructured content.

**Extract Text** asks OpenAI to find specific features in a source string. It
creates one output for every feature, in the same order as the feature list.

## Inputs

| Input                   | Type            | Required | Description                                           |
| ----------------------- | --------------- | -------- | ----------------------------------------------------- |
| **Third-party account** | OpenAI account  | Yes      | The API key used for extraction.                      |
| **Model**               | OpenAI model    | Yes      | Defaults to GPT-5 Mini.                               |
| **Text**                | String          | Yes      | The unstructured source content.                      |
| **Features**            | List of strings | Yes      | One or more unique values describing what to extract. |

Example features for an invoice:

* `vendor name`
* `invoice number`
* `total amount with currency`
* `payment due date in ISO format`

## Outputs

The action creates **Extracted Feature 1**, **Extracted Feature 2**, and later
outputs that line up with the feature inputs. An output is empty when the model
cannot find that feature.

<Warning>
  Do not repeat a feature. Duplicate feature strings cause the action to fail.
</Warning>

<Tip>
  Include the desired format in the feature description. “Due date” may return
  prose; “due date in `YYYY-MM-DD` format” is easier for later actions to use.
</Tip>