Wholesale Catalog Normalizer & Buyer Feed Builder (NEWAPI04)

Normalize a catalog you own or have the right to process into product or brand records for sourcing, Clay, n8n, and spreadsheets. This Actor is a feed processor, not a catalog discovery service. It does not crawl Faire or any marketplace.

Quick start

Run with no input to process three actual food products from the bundled Open Food Facts example. The example shows field mapping and brand/product output. It is not a wholesale feed: wholesale prices, minimum order amounts, MOQ, lead times, locations and shipping availability are unknown and remain null or empty. To process your own data, paste a CSV with headers or JSON array into catalogText, or supply a direct public HTTPS catalogUrl. Set exampleData=false when you want a missing source to be an error.

Example (customer-provided values shown for format illustration, not verified wholesale offers):

{
  "mode": "product",
  "exampleData": false,
  "catalogText": "brand_name,product_name,sku,price,currency,minimum_order_quantity,minimum_order_amount\nExample Brand,Example Cup,C1,8,USD,12,150",
  "format": "csv",
  "maxResults": 100
}

Bundled real example files: data/openfoodfacts_sample.csv and data/openfoodfacts_sample.json. The CSV was derived from three product records retrieved on 2026-10-17 using the official Open Food Facts product API: barcodes 3017620422003, 5449000000996, and 8076800195057. The source URLs are included in each row. The brands field is preserved verbatim as brand_name; it may contain multiple names, and should not be treated as verified manufacturer identity.

Input and rules

Required columns: brand_name, product_name. Optional fields: brand_url, brand_location, website, product_url, sku, category, price, currency, minimum_order_quantity, minimum_order_amount, lead_time_days, shipping_destinations, rating, review_count, tags, source_name, source_license, source_url. Common aliases (vendor, supplier, moq, wholesale_price, ships_to) are accepted. Missing scalar fields are null; missing lists are empty. Invalid rows lacking either required field are skipped, while a feed with only invalid rows fails. Malformed CSV/JSON, source HTTP errors, and invalid row types fail rather than returning an apparently successful empty result.

Choose mode=product or mode=brand. Filter by keyword, brand location, shipping destination, maximum minimum order amount, or maximum lead time, and cap with maxResults. A filter requiring a missing field excludes that record. A valid source with no matching records succeeds with an empty Dataset.

The HTTP reader accepts only direct public HTTPS feeds, at most 5 MB and 50,000 rows, with timeouts and up to three attempts on temporary errors. Do not include credentials in URLs. Only upload or link data you may process and export.

Dataset output

In product mode, one unique product becomes one Dataset item. In brand mode, one unique brand in the submitted filtered feed becomes one Dataset item. observed_product_count counts products observed in this feed; catalog_total_count is null. MOQ is a unit quantity and minimum_order_amount is a monetary threshold. The Actor never infers growth or missing wholesale values.

Export the default Dataset as JSON or CSV from Apify for Clay and n8n. Source attribution columns are retained when supplied. A restart of the same run checks items already saved in its default Dataset and skips them; separate runs each process their own Dataset and are charged independently.

Example data rights

The bundled example is adapted from the Open Food Facts database under Open Database License (ODbL) 1.0 and Database Contents License. Attribution: © Open Food Facts contributors. Changes: selected three products, mapped fields to this Actor's CSV/JSON schema, omitted images and most source attributes. The example and its derivative database are offered under ODbL 1.0; the product text contents retain their applicable source license. You may copy, upload, use commercially and redistribute this example subject to attribution and applicable share-alike obligations. Consult the source's reuse guidance when combining with other databases. This example does not license customer catalogs or third-party artwork.

Pricing

Pay per event: the apify-default-dataset-item event is intended as the only billable event at $0.006 per successfully saved unique default Dataset item ($6 per 1,000 items), whether brand or product. Skipped, rejected, filtered, or duplicate rows are not Dataset items. Each new Actor run is charged independently even if it processes an identical catalog. The live Store pricing section is authoritative. Platform compute and storage charges follow the current Apify plan and the Store listing's usage-inclusion setting. Post-run Dataset storage or retrieval may incur standard Apify charges.

Local development

npm ci && npm test && npm start; docker build -t newapi04 .. To run locally, set APIFY_LOCAL_STORAGE_DIR and create storage/key_value_stores/default/INPUT.json.