> ## Documentation Index > Fetch the complete documentation index at: https://docs.abbyy.com/llms.txt > Use this file to discover all available pages before exploring further. # Data export formats > Reference for export formats and options on the Output activity, covering field values, document text, document images, and error output. Vantage provides export options for document fields, text, and images. ## Export field values (Fields tab) | Format | Export Option | Description | File Name | | :- | :- | :- | :- | | **JSON** | **Values, metadata, and field structure for each document** | Full data extraction results. See [JSON Schema](/vantage/developer/output/json/json-schema) for the output structure. | `.json` | | **JSON** | **Values only** | Field values and rule errors. See [JSON Schema](/vantage/developer/output/json/json-schema) for the output structure. | `_fields.json` | | **CSV** | **Values only** | Field values. | `.csv` | ### CSV output for repeating objects If a document contains repeating objects (repeating fields, repeating groups of fields, tables), Vantage exports them as separate files. The naming scheme is: * Parent CSV file: `.csv` * Child CSV files for repeating objects: `_.csv` * If the child CSV filename would exceed 250 characters, the alternate scheme is used: `_.csv` If a repeating structure is nested within a group, its name appears as `New Group/New Table`. In the child filename, the slash is replaced with an underscore (for example, `New Group_New Table_055fe8c.csv`). ### CSV file format * The field names are written into the first row. * Columns are comma-separated. * Encoding is UTF-8 with BOM. * Empty instances of repeating fields or groups, and empty table rows, are not exported, so the resulting CSV has no empty rows. * Values of **image fields** are empty in the CSV file. ### CSV output in a shared folder When exporting to a shared folder, the parent CSV file is saved in the transaction folder. If the transaction contains multiple documents of the same type, Vantage generates a separate CSV file for each: * A numeric postfix starting from 2 is added to the CSV filename. * A subfolder is created in the transaction folder for child CSV files, named `_` or `_fields_` (if the JSON **Values only** export option is enabled), where N is the sequential document number in the transaction (starting from 2 when there's more than one document). * For repeating objects, the name of the child CSV file is specified in the parent CSV file's field value. ## Export document text (Text tab) | Format | Export Option | Description | File Name | | :- | :- | :- | :- | | **JSON** | **Text only** | A JSON file containing only recognized text; document layout is not preserved. | `_text.json` | | **JSON** | **Preserve document structure** | A JSON file containing recognized text with the document layout preserved. | `_text.json` | | **XML** | **Text only** | An XML file containing only recognized text; document layout is not preserved. | `.xml` | | **XML** | **Preserve document structure** | An XML file containing recognized text with the document layout preserved. | `.xml` | | **ALTOXML** | **Text only** | An ALTO XML file (schema 4.2) containing only recognized text; document layout is not preserved. | `.xml` | | **ALTOXML** | **Preserve document structure** | An ALTO XML file (schema 4.2) containing recognized text with the document layout preserved. | `.xml` | | **TXT** | | A plain text document. The original document structure is preserved using whitespace. | `.txt` | | **DOCX** | **Editable** | An editable Word document which may not look exactly like the original. | `.docx` | | **DOCX** | **Exact** | A non-editable Word document. The original document structure is fully preserved. | `.docx` | | **XLSX** | | An editable Excel document. The original document structure is preserved. | `.xlsx` | | **PPTX** | | An editable PowerPoint document. The original document structure is preserved. | `.pptx` | | **HTML** | | An HTML document that preserves the original document structure. | `.html` | Selecting **Text only** for JSON, XML, or ALTOXML makes export to DOCX, XLSX, and PPTX unavailable. The **Text only** and **Preserve document structure** options cannot differ across JSON, XML, and ALTOXML. Setting one updates the same option for the others. Exported text reflects any field-value changes made by the Manual Review operator during manual review. ### Text output in a shared folder When exporting to a shared folder, all files are saved in the transaction folder. Numeric postfixes starting from 2 are added to filenames when there's more than one document of the same type in the transaction. ## Export document images (Image tab) | Format | Export Option | Description | File Name | | :- | :- | :- | :- | | **PDF** | **PDF/A-3a** | A PDF file with a text layer over the document image. Reflects Manual Review changes. | `.pdf` | | **PDF** | **PDF/A-3b** | A PDF file with a text layer over the document image. Reflects Manual Review changes. | `.pdf` | | **PDF** | **Image-only** | A non-editable PDF in PDF/A-3b standard without a text layer. | `.pdf` | | **TIFF** | | An enhanced image in TIFF format. | `.tiff` | | **JPEG** | **Maximum quality** | An enhanced JPEG image at 95% quality. | `pages/page_.jpg` | | **JPEG** | **Smaller file size** | An enhanced JPEG image at 75% quality, readable while reducing file size. | `pages/page_.jpg` | For each PDF export option, you can choose between **Smaller file size** (default) and **Maximum quality**. Smaller file size uses Mixed Raster Content (MRC) compression, which picks optimal compression separately for text, pictures, and background. ### Image output in a shared folder When exporting to a shared folder, a subfolder is created for each document in the transaction: * The subfolder name is `_`, or `_fields_` if the JSON **Values only** export option is enabled. N is the sequential number of the document in the transaction (starting from 2 when there's more than one document). * Within the subfolder, a `Pages` subfolder stores the JPG files, named `page_.jpg`, where N is the page number. * PDF and TIFF files are saved in the transaction folder. * Numeric postfixes starting from 2 are added to filenames when there's more than one document of the same type. ## General naming scheme Most exported files contain `` in their names, which is one of: * The name of the last **Document** skill applied to the document. * The name of the last **Classification** skill applied, if no Document skill was applied. * `Unknown` if no Document or Classification skills were applied, even though at least one exists in the Process skill flow. If there are multiple output files and export to a shared folder is configured, incrementing numbers are appended to filenames to keep each name unique. ## Error output If a transaction does not complete successfully, Vantage generates an `Error.json` file containing: * Transaction identifier * Transaction status (`Failed`) * Error message * Array of all source file identifiers and names in the transaction Exported data is stored for 2 weeks by default, per the retention policy. ## Related topics * [Output activity](/vantage/documentation/skill-designer/process/output-activity/output-activity) * [Export results to a shared folder](/vantage/documentation/skill-designer/process/output-activity/export-shared-folder) * [Export results to an external system](/vantage/documentation/skill-designer/process/output-activity/export-external-system) * [JSON Schema](/vantage/developer/output/json/json-schema) * [XML Schema](/vantage/developer/output/xml/xml-schema)