RowsInstall

Summary

Extract data from image files using AI vision and imports into a Rows spreadsheet

Inputs

• binaryPropertyName (required) — Name of the binary property that contains the file(s) to upload. Default is 'data'. Files are sent to the API as 'files' parameter. Supported formats: png, jpg, jpeg, webp, pdf, heic, csv, tsv, xls, xlsx. CSV, TSV, XLS, and XLSX require Mode to be set to 'Create'. Max size: 100MB per file, 100MB total.
• collectAllItems — When enabled, collects files from all input items and sends them in a single API request. When disabled, processes each item separately.
• folderId — The folder to use. Choose from the list, or specify an ID using an expression.
• appId — The spreadsheet to use. Choose from the list, or specify an ID using an expression. Sent to API as 'app_id'. Filtered by Folder ID if provided.
• tableId — The table to use. Choose from the list, or specify an ID using an expression. Required when Spreadsheet ID is provided.
• mode — Processing mode that affects the response format. Options: Create (create new data from the image), Read (read and extract data from the image), Read Simplified (read and extract simplified data from the image).
• merge — Whether to merge with existing data.
• instructions — Custom instructions for data extraction (e.g., 'Extract all data from this image').

Output shape

Returns the vision extraction result as a JSON object attached to each input item.

If collectAllItems is true (default), all files from all input items are sent in a single API request. If false, each item is processed separately with its own API request.

Examples

Example 1: Import a single image file into a Rows table with AI vision extraction in 'Create' mode.

Set binaryPropertyName to the binary property containing the image, leave collectAllItems at default (true), optionally specify folderId/appId/tableId, set mode to 'create' (default), and optionally provide custom instructions.