Bring the export.
Keep the evidence.

Start with a fresh Shopify product export.

Use a complete CSV for the products being repaired. Keep handles, variant SKUs, options and the original columns. The preflight reports the fields it recognizes and anything that makes the file ineligible. A title match alone cannot identify a variant.

Current and older Shopify header names can differ. Do not rename unfamiliar columns by guesswork. Use the checker’s explicit mapping and consult Shopify’s current CSV guide before any import.

Download a fictional sample CSV. For a callable interface, read the agent workflow.

Supply exact-item sources.

Use documents that identify the precise SKU and state the facts you want corrected. A machine-readable product sheet, structured supplier file or approved specification record is more useful than a broad product-family brochure. Scanned or image-only PDFs are unsupported.

Specstead checks proposed evidence against the uploaded text. It cannot establish that a supplier document is genuine, current or complete. Product dimensions, packaging dimensions and installation requirements must keep their separate meanings.

Review a correction, keep the original, or resolve a conflict.

Each proposal shows the original value, proposed option, affected SKUs and source passage. Accept only the supported value you intend to use. You can retain the original. Conflicting evidence requires a deliberate decision; unsupported facts remain on hold.

Supported existing columns are Title, scalar product metafield values, weight in grams and variant image URL. Description HTML is not repaired. Identity, handles, options, prices, inventory and other fields outside the accepted repair stay unchanged. A shared product-level field needs evidence that agrees across the affected variants.

The CSV does not include metafield type definitions. Specstead does not validate Shopify field types or guarantee import acceptance. Check that the reviewed scalar value is valid for the existing field; lists, structured JSON and resource references are outside this repair’s scope.

Image URLs are matched to supplied evidence; the service does not fetch or visually inspect those images. Keep all existing variants for each included handle in your export. The checker cannot discover a variant omitted from the file.

Buy the reviewed version. Import when you are ready.

The paid ZIP is tied to the accepted revision. It contains catalog-corrected.csv, changes-review-only.csv, evidence.json and READ-ME.txt. Checkout is available only after every proposal has an explicit choice, at least one real correction is accepted, and the payment service is available. The merchant performs the import; Specstead has no store-write access.

The corrected catalog is a full snapshot. It preserves the original values outside the accepted changes. If someone changes a price, inventory value or other field in your store after the export, importing this snapshot can restore that older value. Get a fresh export and compare the current record before importing.

Keep the original export and review Shopify’s import settings. After importing, export the affected products again and upload that file for comparison. The purchase includes up to ten rechecks during the active paid-workspace period. A recheck reports selected matches and differences in supplied data; it does not inspect the live store or guarantee the platform accepted every field.

See the edge cases.

Read what we tested for the synthetic checks, competitor research and limits of the evidence.

• Use the supplier PDF to repair the right CSV cells.

  How to prepare an existing Shopify product family and readable supplier files for an evidence-backed specification repair.

• Different variant facts need variant-level fields.

  Why different sibling specifications cannot be repaired into a shared product metafield, and where native Shopify product CSV reaches its limit.

• Product weight and shipping weight are different facts.

  A worked review of net product weight, packaged weight and grams in a Shopify product CSV.

• Keep the right facts with the right variant.

  A worked example of product-family identity: why a matching title cannot prove that a specification belongs to the right Shopify variant.

• When the source files disagree, keep the conflict visible.

  A product-data conflict example showing source identity, exact evidence, measurement basis and why an unresolved fact should remain unchanged.

• Review the import, not just the corrected cells.

  Understand why a Shopify product CSV repair should preserve the original snapshot, distinguish blank and omitted values, and keep identity and commerce fields unchanged.

Private files stay out of public discovery.

The free preflight does not persist a catalog. An unpaid workspace expires three days after creation; a paid workspace expires 30 days after its first confirmed payment. The workspace shows the exact expiry. Expired workspaces become inaccessible and are removed by the service’s cleanup process. Save purchased files before access expires.

Private analysis stores uploaded inputs and can send source text to the configured AI provider. Each workspace has up to three analysis attempts, subject to service-capacity limits. Read the data policy before uploading confidential material.

Your private workspace link and access token grant access to the job. Keep them private. Public examples are fictional and contain no customer files.