Exporting Data

Export product content from merchi.ai as CSV or JSON for use in other systems, offline review, or data pipelines. The export panel gives you control over what’s included, how it’s formatted, and how images and attributes are structured.

How to Export

1. Go to the All Products View (/grid)
2. Apply any filters you need — by taxonomy, status, date range, or search
3. Click the Download button to open the Export panel
4. Choose your format (CSV or JSON), configure options, and click Export

Choosing What to Export

The Export panel offers two modes:

Export selected rows

Check product rows in the grid first, then open the Export panel. The selected rows option will show how many products you have checked. This is useful for handpicking a specific set.

Export all matching filters

Use the all matching filters option to export every product that matches your current grid filters — up to 2,000 products — without selecting each one manually. The panel shows the live count so you know exactly how many products will be included before you click.

If your filters match more than 2,000 products, the export will not run. Narrow your filters (by date range, taxonomy, or status) until the count is within the limit.

Available Filters for “All Matching”

The export uses all filters currently applied in the grid:

• Locale — which language/market’s content to export
• Category / Taxonomy — filter by product category
• Workflow status — Approved, In Review, Pending, Rejected
• Updated date — products updated within a date range
• Text search — title or description contains a term

The Export panel also adds a Created date filter not available on the main grid, for filtering products by when they were first processed.

CSV Options

Image layout

• One column per image — each image gets its own column with position, URL, and alt text
• One row per image — a separate row is created for each image, with product data repeated
• All images in one cell — all image URLs combined in a single cell, separated by a delimiter you choose

Attribute layout

• Wide — each attribute appears in its own column (best for spreadsheet analysis)
• Single column — all attributes combined in one cell (best for compact exports)

Delimiter

Choose comma, semicolon, or tab as the column separator. Semicolons are common for European locales where commas appear in numbers.

Boolean format

Choose how true/false values appear: true/false, Yes/No, or 1/0.

Empty value handling

Choose what happens when a field has no data: omit the field entirely, output null, or output an empty string.

Header row

Toggle off if the receiving system does not expect column headers.

JSON Options

Structure

• Nested (default) — attributes appear as an array of objects [{key, value}], matching how the data is stored
• Flattened — keys use dot-path notation (attributes.0.key, attributes.0.value), useful for ETL tools that expect flat key-value JSON

Output shape

• Array — a single JSON array [{...}, {...}] — standard for most uses
• NDJSON — one product object per line, no wrapping array — preferred for streaming pipelines, BigQuery, and log-style ingestion

Pretty print

Toggle on for human-readable indented JSON (e.g. for reviewing in an editor). Toggle off for compact output when piping into another system.

Saved Configs

Once you’ve configured an export the way you need it, save it with a name using the Saved configs section at the bottom of the panel. Next time, load it from the dropdown and all your settings are restored instantly.

You can save separate default configs for CSV and JSON. The config marked as default will be pre-loaded when you open the panel.

Language Selection

The export uses the locale you select in the panel (defaulting to the locale currently active in the grid). To export in multiple languages, run a separate export for each locale.

Practical Examples

Export approved products for a platform import

1. Filter by Approved workflow status
2. Open the Export panel
3. Choose All matching filters — confirm the count
4. Select CSV, set your image and attribute layout to match the target platform
5. Download and map columns as needed

Export recent additions for review

1. Open the Export panel
2. Use the Created date filter to select products from the last 7 days
3. Export as CSV and share with your team

Export to a data pipeline

1. Set up your filters to match the product set you want
2. Choose JSON with NDJSON output shape and Pretty print off
3. The resulting file is ready to stream directly into BigQuery, a log pipeline, or a custom ingestion tool

Limits

• Maximum export size: 2,000 products per export (filter to stay within this)
• One locale per export
• Image files are linked by full URL — the files themselves are not embedded