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

# Create a new extraction task

## Input Payload

* `url` (string): The URL from which data will be extracted. It should point to the specific page containing the data of interest.
* `query` (string): The query parameter allows you to define the extraction query to specify the data that needs to be extracted from the page.
* `columns` (string): A comma-separated list of column names that acts as an interface to structure the extracted data.
* `limit` (integer): Set a limit to the number of rows to extract
* `format` (string): The output format how you'd like to retreive the data (possible values: `csv`, `markdown`, `json`). Defaults to `json`

## Output Payload

When the extraction task is successfully initiated, the API responds with a `200` status code, indicating a successful request. The response includes a payload with several important pieces of information. Among these, the `data.id` field is particularly vital.

### Saving the Run ID

```javascript
// Response example with a status code of 200
{
  "success": true,
  "data": {
    "id": "your-unique-run-id",
    // Additional fields...
  },
  // Other response properties...
}
```

<Note>
  Ensure to save the `data.id` after receiving the response.
  You will use this unique ID to periodically keep polling the GET endpoint to receive the extracted data.
</Note>