Shopify Product CSV Validator & Cleaner

Checks your Shopify product CSV before you import it and returns every variant's verdict with its errors and warnings, one row per issue and a summary, plus a cleaned CSV with the safe fixes; failed and unchanged rows are never charged.

Paste your product CSV (or pick the key-value store that holds it) and see, line by line, what Shopify's import would refuse or misread: unknown or misspelt columns, missing titles, bad URL handles, broken variant options, duplicate SKUs and barcodes, GTIN/EAN/UPC check digits, prices with currency symbols or decimal commas, compare-at prices below the price, invalid image URLs, wrong true/false and status values, weights with units. Built for Shopify merchants and agencies, dropshippers with supplier feeds, migrations from other platforms, and AI agents that prepare listings. Not affiliated with or endorsed by Shopify.

What it does

• Checks the whole file against Shopify's published product CSV format: the current column names (URL handle, Description, SKU, Barcodes, Option1 name …) and the older ones Shopify still accepts (Handle, Body (HTML), Variant SKU, Variant Price, Image Src …), market, Google Shopping and metafield columns.
• One row per variant with its verdict (ready, warnings or errors), a qualityScore, its issue codes and each issue's line, column, message and suggested value. A product's own issues (title, handle, images, tags) are listed on its first variant.
• One free row per issue (rowType: "issue"), handy in a spreadsheet, and one free summary row per file: counts, importReady, feedQualityScore, the issue counts and the unknown columns.
• A cleaned CSV (CLEANED_CSV in the run's key-value store): your file with only the safe fixes applied (trimmed cells, column name letter case, Yes → true, Active → active, $29.99 → 29.99, 1,299.00 → 1299.00, 10.0 → 10, UTF-8 with LF line ends). It never invents or changes a fact: no title, brand, barcode, price, weight, handle or image is ever made up or changed; those get a suggestion only.
• Watch mode (give a watch list name in stateName): upload each new version of your feed and get only the variants that are new or whose issues changed, plus a price_jump warning when a price moved by priceJumpPercent or more. A watch run with nothing new returns exactly one free no_data row that says so.
• Nothing is fetched: image URLs are checked as text only; the actor sends no request to Shopify or to your images.
• You are never charged for failed results: failed, no_data, issue, summary and unchanged rows are free.

Quick start

Paste your CSV (header line first) in Product CSV (text), or as JSON:

{ "csvText": "URL handle,Title,Option1 name,Option1 value,SKU,Price\nmug,Stoneware Mug,Title,Default Title,MUG-350,12.50\n", "currency": "USD" }

A large file in your own key-value store (upload it with the Apify API or another Actor), checked every week in watch mode:

{ "keyValueStoreId": "aBcDeFgHiJkLmNoPq", "recordKey": "products.csv", "stateName": "my-store", "currency": "EUR", "decimalComma": true }

Input

Field	Type	Default	Description
csvText	text	— (prefilled with a small sample)	Your Shopify product CSV, header line first; up to about 9 MB
keyValueStoreId	key-value store (picker)	—	Your store that holds the CSV as a record; the picker gives this run read access to it only
recordKey	string	—	The record's key, e.g. products.csv
mode	watch or export	—	Empty: watch when stateName is given, otherwise export. An explicit mode wins
stateName	string	—	Watch list name. Watch mode without a name uses the list default
currency	ISO 4217 code	—	Your store's currency, checked and shown beside every price
decimalComma	boolean	false	Prices use a decimal comma (29,99): read and fixed as 29.99
targetMarginPercent	integer 0–99	—	Warn when (price − cost per item) / price is under this percent
priceJumpPercent	integer 1–10,000	50	Watch: a price that moved this much since the last run is a price_jump warning
issueRows	boolean	true	Also write one free row per issue
writeCleanedCsv	boolean	true	Save CLEANED_CSV with the safe fixes
maxItems	integer 1–200,000	100000	Most charged variant rows per run; in watch mode the rest come in the next run

Both csvText and a store record may be given: each file is checked on its own (the second cleaned file is CLEANED_CSV_2). Changing decimalComma or targetMarginPercent starts a new watch baseline.

Output

A variant row:

{
    "status": "ok",
    "attempts": 0,
    "error": null,
    "input": "csvText",
    "rowType": "variant",
    "mode": "export",
    "watchList": null,
    "noDataReason": null,
    "sourceRow": 3,
    "handle": "pet-gps-tracker",
    "sku": "GPS-RED",
    "options": "Color: Red",
    "price": 29.99,
    "compareAtPrice": 24.99,
    "costPerItem": null,
    "currency": "USD",
    "barcodes": ["8712345678902"],
    "verdict": "errors",
    "qualityScore": 65,
    "errorCount": 1,
    "warningCount": 2,
    "issueCodes": ["gtin_check_digit", "price_currency_symbol", "compare_at_not_higher"],
    "issues": [
        { "severity": "warning", "code": "gtin_check_digit", "column": "Barcodes", "sourceRow": 3, "message": "This looks like a GTIN/EAN/UPC but its check digit is wrong: Shopify keeps it as a custom barcode, and Google or Meta will reject it. Confirm it with the manufacturer.", "suggestedValue": null, "fixApplied": false },
        { "severity": "error", "code": "price_currency_symbol", "column": "Price", "sourceRow": 3, "message": "Price holds a currency symbol or code; Shopify wants the number only (9.99).", "suggestedValue": "29.99", "fixApplied": true },
        { "severity": "warning", "code": "compare_at_not_higher", "column": "Compare-at price", "sourceRow": 3, "message": "The compare-at price (24.99) is not above the price (29.99), so no discount shows.", "suggestedValue": null, "fixApplied": false }
    ],
    "changeType": null,
    "previousIssueCodes": null,
    "priceChangePercent": null,
    "url": "https://console.apify.com/storage/key-value-stores/aBcDeFgHiJkLmNoPq",
    "source": "Your own product CSV, checked against Shopify's published product CSV format. Not affiliated with or endorsed by Shopify.",
    "scrapedAt": "2026-10-05T09:00:00.000Z"
}

(The issue and summary columns of a variant row are null; the dataset's Issues and Summary views show those rows.)

Status	Meaning	Charged?
ok	A checked variant (rowType: "variant"), an issue row or a summary row	Variant rows: export yes; watch baseline, new and changed yes. Issue and summary rows: no
no_data	Only a header line, image lines only for a handle, no such record, or nothing new since the last run	No
failed	The file can't be read (not a Shopify CSV, not comma-separated, an unclosed quote), a line has the wrong number of values, or the record can't be read	No

The key-value store holds CLEANED_CSV, OUTPUT (each file's summary) and RUN_REPORT (counts, charged and free rows, stop reason).

Issue codes

Errors are what Shopify's import refuses or misreads; warnings import but are likely wrong. ✓ = fixed in CLEANED_CSV.

Code	Severity	What it means
unknown_column	error	A column Shopify's product CSV doesn't have (only Collection may be added)
column_name_case	warning ✓	A known column written with other letter case or spacing
duplicate_column, unnamed_column	error	Two columns for the same field; a column with no name
barcode_columns_conflict	error	Both Barcode and Barcodes (Shopify refuses the file)
variant_metafield_column	error	Variant metafields aren't supported in the product CSV
missing_title_column	error	No Title column
mixed_column_formats, line_endings ✓, not_utf8	warning	Old and new column names mixed; CRLF line ends; text that isn't UTF-8
title_missing, title_too_long	error	No Title on the product's first line; over 255 characters
handle_invalid	error	A URL handle with spaces or other characters (suggestion only)
handle_missing, handle_not_contiguous	warning	No handle (Shopify makes one); a product's lines not together
duplicate_handle	error	Two products share a URL handle
title_on_variant_line	warning	A later line has another Title (Shopify reads the first)
status_missing	error	The Status column is present but empty
too_many_tags, too_many_images, too_many_barcodes	error	Over 250 tags, 250 images, 20 barcodes
seo_title_too_long, seo_description_too_long	warning	Over 70 / 320 characters
alt_text_too_long, collection_too_long	error	Over 512 / 255 characters
image_url_invalid	error	Not an https:// or http:// image address (checked as text)
image_url_not_https, image_file_name_suffix, duplicate_image	warning	http://; a _thumb/_small/_medium file name; the same image twice
option_value_missing, option_name_missing, option_name_changed	error	Option names and values that don't match
duplicate_option_combination	error	Two variants with the same options
variants_without_options, default_title_with_variants	error	Several variants but no option; Default Title with several variants
duplicate_sku, whitespace ✓	warning	A SKU used twice; spaces around a handle, SKU or barcode
sku_missing_custom_fulfillment	error	A custom fulfillment service without a SKU
price_currency_symbol ✓, price_thousands_separator ✓	error	$29.99, 1,299.00
price_decimal_comma	error	29,99 (✓ with decimalComma)
price_invalid, price_negative	error	Not a number (free, 29 dollars); below 0
price_missing, price_zero, price_decimals	warning	No price (Shopify sets 0.00); 0; more than 2 decimals
compare_at_not_higher, cost_above_price, margin_below_target	warning	No discount shows; a loss on each sale; under your margin target
invalid_boolean	error	Not true/false (✓ for yes/no/1/0)
invalid_enum	error	Status, inventory policy, tracker, units, fulfillment service (✓ for letter case)
invalid_integer, invalid_number, weight_not_whole_grams	error	Not a whole number (✓ for 10.0); not a number; grams with a unit or decimals (suggested in grams)
packed_size_partial	error	Only some of the four packed product columns
related_value_missing	warning	Inventory tracked but the out-of-stock policy blank
gtin_check_digit, gtin_length, isbn_check_digit	error / warning	A wrong GS1 check digit or length: an error for a typed barcode (gtin:, upc:, ean:), a warning for an untyped one (Shopify keeps it as a custom barcode, Google and Meta reject it)
barcode_scientific_notation	error	A spreadsheet turned the barcode into 8.71E+12
gtin_leading_zero_lost, duplicate_barcode	warning	An 11-digit UPC (leading zero dropped); one barcode on two variants
price_jump	warning	Watch: the price moved by priceJumpPercent or more since the last run

Output fields

Field	Type	Description
status	string (or null)	ok = a checked variant (charged in export mode, and in watch mode when it is a baseline, new or changed variant), an issue row or a summary row (both free); no_data = nothing to check (only a header line, image lines only for a handle, no such record, or nothing new since the last run) (free); failed = the file or a line could not be read (free)
attempts	integer (or null)	Reads made for this row: 0 for text in the input, 1 for a key-value store record
error	string (or null)	Why a row is no_data or failed; null on ok rows
input	string (or null)	Which file the row is about: csvText, or the key-value store ID and record key
rowType	string (or null)	variant = one variant line (its verdict and issues); issue = one issue (free); summary = the file (free)
mode	string (or null)	watch or export
watchList	string (or null)	Watch list name (watch mode); null in export mode
noDataReason	string (or null)	Why a row is no_data: empty_file (only a header line), images_only (image lines only for a handle), nothing_new (watch: no change since the last run), not_found (no such record); null otherwise
sourceRow	integer (or null)	The line in your file (1 = the header line), as your spreadsheet numbers it
handle	string (or null)	The product's URL handle
sku	string (or null)	The variant's SKU
options	string (or null)	The variant's options, e.g. Color: Red / Size: S
price	number (or null)	The price as Shopify would read it (null when it is missing or unreadable)
compareAtPrice	number (or null)	The compare-at price as Shopify would read it
costPerItem	number (or null)	The cost per item as Shopify would read it
currency	string (or null)	Your store currency (the currency input, ISO 4217), shown beside the prices; null when not given
barcodes	array (or null)	The variant's barcodes (from Barcodes or the older Barcode column)
verdict	string (or null)	ready (no issue), warnings (warnings only) or errors (at least one error: Shopify would refuse or misread something)
qualityScore	integer (or null)	100 minus 25 per error and 5 per warning (0 to 100)
errorCount	integer (or null)	The variant's errors (its product's issues count on its first variant)
warningCount	integer (or null)	The variant's warnings
issueCodes	array (or null)	The codes of the variant's issues (see the README's issue table)
issues	array (or null)	Each issue: severity, code, column, line, message, suggested value and whether CLEANED_CSV applies it
changeType	string (or null)	Watch mode: baseline (first run of the list), new (a variant the list did not have) or changed (its issues changed, or its price jumped); null in export mode
previousIssueCodes	array (or null)	Watch mode: the variant's issue codes at the last run
priceChangePercent	number (or null)	Watch mode: how much the price moved since the last run
severity	string (or null)	Issue rows: error (Shopify would refuse or misread it) or warning (it imports, but something is likely wrong)
issueCode	string (or null)	Issue rows: the issue code (see the README)
column	string (or null)	Issue rows: the column as named in your header line; null for a whole-line or whole-file issue
value	string (or null)	Issue rows: the cell's value, only for identifiers, numbers and codes (never titles, descriptions or vendors)
message	string (or null)	Issue rows: what is wrong, in one sentence
suggestedValue	string (or null)	Issue rows: the value to use instead, when one is clear
fixApplied	boolean (or null)	Issue rows: true when CLEANED_CSV already has the suggested value (a safe formatting fix)
relatedRows	array (or null)	Issue rows: the other line(s) involved (the first duplicate, the product line)
rowsRead	integer (or null)	Summary: data lines read (the header not counted)
productCount	integer (or null)	Summary: products (URL handles)
variantCount	integer (or null)	Summary: variant lines checked
imageRowCount	integer (or null)	Summary: lines that only add images
variantsReady	integer (or null)	Summary: variants with no issue
variantsWithWarnings	integer (or null)	Summary: variants with warnings only
variantsWithErrors	integer (or null)	Summary: variants with at least one error
errorTotal	integer (or null)	Summary: all errors (variants, products and file)
warningTotal	integer (or null)	Summary: all warnings
fileIssueCount	integer (or null)	Summary: issues of the header or the file itself
importReady	boolean (or null)	Summary: true when the file has no error and every line could be read
feedQualityScore	integer (or null)	Summary: the variants' average quality score, minus 10 per file error and 5 per unreadable line (0 to 100)
csvFormat	string (or null)	Summary: current (today's column names), legacy (the older names Shopify still accepts) or mixed
unknownColumns	array (or null)	Summary: columns that are not in Shopify's product CSV
issueCounts	array (or null)	Summary: how many of each issue, most frequent first
cleanedCsvKey	string (or null)	Summary: the record in the run's key-value store that holds the cleaned CSV (null when writeCleanedCsv is off)
unchangedCount	integer (or null)	Watch mode summary: variants, lines and file issues unchanged since the last run (not returned)
removedCount	integer (or null)	Watch mode summary: variants in the last run that are not in this file
url	string (or null)	The run's key-value store in Apify Console, which holds CLEANED_CSV and OUTPUT
source	string (or null)	Where the data comes from
scrapedAt	string (or null)	When the row was written (ISO 8601, UTC)

Pricing

Pay per event: one variant-check event per checked variant (export: every variant; watch: baseline, new and changed variants).

Plan	Price per variant	Per 1,000 variants
Free (no discount)	$0.0015	$1.50
Bronze	$0.0013	$1.30
Silver	$0.00115	$1.15
Gold (and Platinum, Diamond)	$0.0010	$1.00

• Never charged: failed rows (a file or line that can't be read), no_data rows, issue rows, summary rows, image lines and unchanged variants. You are never charged for failed results.
• Apify also charges its small per-run start event. There are no usage fees on top, so the actor is eligible for x402 agent payments.

Use it from AI agents

One clear main input, csvText (the CSV as text); every variant row has status, verdict, issueCodes and issues, and the summary row has importReady. Call it through the Apify API, the Apify MCP server (mouadapi/shopify-product-feed-validator) or x402 agentic payments. Copy-paste call (your Apify token in place of YOUR_APIFY_TOKEN):

$curl -X POST "https://api.apify.com/v2/acts/mouadapi~shopify-product-feed-validator/run-sync-get-dataset-items?token=YOUR_APIFY_TOKEN" -H "Content-Type: application/json" -d '{"csvText": "URL handle,Title,Option1 name,Option1 value,SKU,Barcodes,Price\u000amug,Stoneware Mug,Title,Default Title,MUG-350,4006381333931,12.50\u000a", "currency": "USD"}'

import { ApifyClient } from 'apify-client';
import { readFileSync } from 'node:fs';

const client = new ApifyClient({ token: 'YOUR_APIFY_TOKEN' });
const run = await client.actor('mouadapi/shopify-product-feed-validator').call({ csvText: readFileSync('products.csv', 'utf8'), currency: 'USD' });
const { items } = await client.dataset(run.defaultDatasetId).listItems();
const notReady = items.filter((r) => r.rowType === 'variant' && r.verdict === 'errors');

Limits

• Shopify's product CSV only (V1). Google Merchant Center and Meta catalog profiles are planned.
• Up to 200,000 lines and about 9 MB of text in csvText; a larger file goes in a key-value store record.
• Image URLs are checked as text: an address that is well formed but doesn't exist is not found (the actor fetches no URL).
• Barcodes are checked by their GS1 check digit and length only (no barcode database): a valid number can still belong to another product.
• Watch mode keeps one entry per variant in your own named key-value store; variants removed from the feed are counted in removedCount.

Known issues

• Shopify's importer can change its messages and its column names (it keeps older names working). The checks follow its published format; an import can still fail for reasons the file doesn't show (an image Shopify can't download, a custom fulfillment service that isn't set up in your admin, a product category Shopify doesn't know).
• Option values, SKUs, barcodes, prices and handles come back in the rows; titles, descriptions, vendors, tags and other free text never do (they stay in your file and in CLEANED_CSV).

FAQ

Does it change my file? No. Your input is never changed; CLEANED_CSV is a separate copy with formatting fixes only, and every fix is listed on its issue (fixApplied: true).

Why is a price with a comma an error and not fixed? 29,99 and 1,299 look alike; set decimalComma: true if your prices use a decimal comma, and the cleaned CSV fixes them.

Why did a watch run return one no_data row? Nothing changed since the last run of the watch list; the row says how many were unchanged. It is free.

Data and licence

• Input: your own file. It is read only in this run (and, in watch mode, its variants' issue codes and prices are kept in your own watch list store).
• The rules follow Shopify's published product CSV format (https://help.shopify.com/en/manual/products/import-export/using-csv). This actor is not affiliated with or endorsed by Shopify.