---
title: "Data Value Extractor"
slug: "data-value-extractor"
updated: 2025-09-15T09:21:49Z
published: 2025-09-15T09:21:49Z
canonical: "knowledge-base.rossum.ai/data-value-extractor"
---

> ## Documentation Index
> Fetch the complete documentation index at: https://knowledge-base.rossum.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Data Value Extractor

The Data Value Extractor serves to extract data from a document that is linked in annotation's metadata. The main use case is to process data from [REST API Export](/help/docs/rest-api-export) as a part of the [Export Pipeline](/help/docs/export-pipelines).

## Installation

1. Login to your Rossum account.
2. Navigate to **Extensions** → **My extensions**.
3. Click on **Create extension**.
4. Fill the following fields:
  1. Name: `Data value extractor`
  2. Trigger events: `Export`
  3. Extension type: `Webhook`
  4. URL (see below)
5. In "Advanced settings" select **Token owner** (should have Admin access)
6. Click **Create the webhook**.

| Environment | Webhook URL |
| --- | --- |
| EU1 Ireland | [https://elis.data-value-extractor.rossum-ext.app/](https://elis.data-value-extractor.rossum-ext.app/) |
| EU2 Frankfurt | [https://shared-eu2.data-value-extractor.rossum-ext.app/](https://shared-eu2.data-value-extractor.rossum-ext.app/) |
| US east coast | [https://us.data-value-extractor.rossum-ext.app/](https://us.data-value-extractor.rossum-ext.app/) |
| Japan Tokyo | [https://shared-jp.data-value-extractor.rossum-ext.app/](https://shared-jp.data-value-extractor.rossum-ext.app/) |

## Available configuration options

Simple extraction example.

```json
{
  "extract": [
    {
      "format": "json",
      "source_reference_key": "ifs_export_reply_payload",
      "extract_rules": [
        {
          "value_path": "MessageId[0].value",
          "target_schema_id": "ifs_reply_message_id"
        }
      ]
    }
  ]
}
```

More complex configuration example using extraction from two different `source_reference_key` and two `extract_rules` in the second one. There is also the `condition` used, which is reference to a document ID in the annotation which triggers the execution of the extraction.

```json
{
  "extract": [
    {
      "format": "json",
      "extract_rules": [
        {
          "value_path": "doc_id",
          "target_schema_id": "erp_doc_id"
        }
      ],
      "source_reference_key": "api_xml_export_reply_payload"
    },
    {
      "format": "json",
      "condition": "@{api_gate}",
      "extract_rules": [
        {
          "value_path": "status_code",
          "target_schema_id": "erp_api_status_code"
        },
        {
          "value_path": "headers.etag",
          "target_schema_id": "erp_api_etag"
        }
      ],
      "source_reference_key": "api_xml_export_reply_headers"
    }
  ]
}
```

## Parameters

### Extract Object

The extract object consists of the following parameters:

| Attribute | Type | Description |
| --- | --- | --- |
| `format` | str | File format. Currently, only `json` value is supported. |
| `condition` | str | Reference to `annotation.content` `schema_id` that holds evaluated value. When it's empty or "false" (case insensitive), this section won't be evaluated. Otherwise, it will proceed. The condition follows the [JSON templating](/help/docs/json-templating) syntax e.g. `"condition": "@{api_gate}"` |
| `source_reference_key` | str | Relation key into metadata for source document. |
| `extract_rules` | object | Rules to update annotation's content. |

The `extract_rules` object defines how values are extracted and stored:

| Attribute | Type | Description |
| --- | --- | --- |
| `value_path` | str | Query to get the value from the referred document. In case of `format=json`, it should be in `jmespath` syntax. |
| `target_schema_id` | str | Annotation's `schema_id` to be updated. |

> [!WARNING]
> 💬 YOUR FEEDBACK MATTERS!
> 
> Help us keep this page accurate and useful. Select **Yes** or **No** below, then use the feedback form to propose a correction, ask for clarification, or request a new article.