This explains how to use the POST /pipelines endpoint to extract data and have it returned in Markdown format instead of JSON. All of the functionality remains the same, ie you can extract from multiple sources, use multiple schemas, etc, but the data will be returned as markdown.