curl --location --request POST 'https://api-prod.extend.app/workflow_runs/wr_123/outputs/dpr_123' \
--header 'Content-Type: application/json' \
--header 'Authorization: Bearer <API_TOKEN>' \
--data '{
    "reviewedOutput": {
      // Output object - see “Processor output types” for details
    }
}'
If you are using our Human-in-the-loop workflow review, then we already will be collecting your operator submitted corrections. However, if you are receiving data via the API without human review, there could be incorrect outputs that you would like to correct for future usage in evaluation and tuning within the Extend platform. This endpoint allows you to submit corrected outputs for a WorkflowRun, by providing the correct output for a given output ID. The output ID, would be found in a given entry within the outputs arrays of a WorkflowRun payload. The ID would look something like dpr_gwkZZNRrPgkjcq0y-***. When using this endpoint, you will need to provide both the run id and the output id as parameters, and the corrected reviewedOutput as the body of the request. See below for more details.

​
Parameters

​
workflowRunId
string
The ID of the WorkflowRun that included this initial output.
​
outputId
string
The ID of the output that these corrections are for.

​
Body

​
reviewedOutput
object
required
The corrected output of the processor when run against the file.This should be a JSON object conforming to the output type schema of the given processor.If this is an extraction result, you can include all fields, or just the ones that were corrected, our system will handle merges/dedupes. However, if you do include a field, we assume the value included in the final reviewed value.
curl --location --request POST 'https://api-prod.extend.app/workflow_runs/wr_123/outputs/dpr_123' \
--header 'Content-Type: application/json' \
--header 'Authorization: Bearer <API_TOKEN>' \
--data '{
    "reviewedOutput": {
      // Output object - see “Processor output types” for details
    }
}'