The alternative to rewriting Shopify products one by one is a CSV bulk job, with a person reviewing the drafts before they are published. On Squadexa AI, that job starts at Basic: $9 a month, up to 10 products per CSV. Pro is $49 a month, up to 50 products per CSV, plus a Shopify import CSV download. Sidekick in the desktop admin remains the alternative when the pile is only a handful of products. Both can be true in the same store.
A bulk job fills the app’s product library. It does not replace live descriptions by itself. The operations win is that nobody opens fifty product pages to paste fifty paragraphs. The operations risk is that fifty drafts get published because the file finished. Keep the review. Confirm a pre-regeneration snapshot exists before you count on revert for any row that came from the CSV.
What one-by-one actually costs in time
Opening a product, reading the old text, writing or pasting a new description, checking the SEO fields, and saving is a short task. Repeated across a catalog, it becomes the week. This article does not convert that week into an hourly rate or a dollar savings. The useful comparison is the motion: one product admin page per SKU, versus one spreadsheet and a review queue.
The manual motion is still the right one for a product with a tricky warning, a fit note, or a certificate. It is the wrong default for a hundred title-only imports that share a simple structure. Put the simple rows in the sheet. Leave the tricky rows for a person in the product admin, or for regenerate-existing where you can compare one product at a time.
- One-by-one: a handful of products, or any product whose wording is already careful.
- CSV: many thin listings that share a fact pattern you can put in a sheet.
- Review either way: someone reads measurements and claims before the text is public.
Basic versus Pro
Choose the plan from the job size. Basic stops at 10 products in a CSV job. A catalog of 30 thin products is three Basic jobs, with a review after each, or it is one reason to look at Pro. Pro stops at 50 products in a job. A catalog larger than 50 is multiple Pro jobs. There is no Enterprise tier above that in the catalog used here, and there is no word cap to stretch the job. Free, at $0, does not run the CSV.
The Squadexa page includes a starter template. Use it, and do not guess column names. Each row needs the product name and the keywords you can verify: material, size, who it is for, and one use. A blank keyword cell produces a vague draft that the reviewer will have to rebuild, which puts you back in one-by-one editing.
Progress on the job updates every few seconds and survives a refresh. When it finishes, the records are in the library with status badges for created, updated, not created, and not updated. That queue is the review column in practice. Work it from the top. If the first rows share an invented claim, fix the sheet before the next job.
Review is still required
Bulk changes who types the first draft. It does not change who is responsible for the live text. The generation is keyword-aware and editable. It does not document a fact check. Measurements, certifications, and competitor comparisons stay with the reviewer. Ordinary sentences about what the product is and who it is for can pass after a read.
Save for later holds an edited record without writing to Shopify. Use it as the queue between the operator who ran the CSV and the person who approves the words. The catalog does not document a formal approval status. The badge “not updated” is the closest signal that the live product is unchanged. Publish when the read is done, and publish one field when the body should stay and only the meta description should move.
- Read the description, the SEO title, the meta description, and the tags as separate checks.
- Delete tags that do not match the labels your collections already use.
- Leave price and status alone unless this project was meant to change them.
- Stop the queue if the same unsupported claim appears on several rows.
The import file on Pro
Pro’s CSV download maps generated products to Shopify’s product-import schema, including handle, body HTML, variant price, and SEO fields. That file is the alternative to clicking publish inside the app, for a team that already imports products from a spreadsheet. Open the file. Check handles so you update the right product. Check prices so a copy project does not change what the item costs. Then import through Shopify when the sheet is clean.
Basic does not include that download in the plan statement used here. On Basic, publishing happens from the library, 10 generated products at a time. A create in the app is followed by a variant price update. An update can send title, description, tags, SEO fields, and status. Sample one record end to end before you publish the other nine.
When to keep Sidekick
Shopify Help documents Sidekick for a product description in the admin, on a desktop computer. Use it when the week’s work is a few products, or when a single awkward SKU should not wait for the next CSV. The help page says you edit the suggestion before you save, and that you are responsible for accuracy. A handful of Sidekick drafts plus a CSV for the long tail is a coherent operation. You do not have to pick one tool for every product.
Revert is the recovery for a regenerate, not a blanket undo for the spreadsheet. The app snapshots the current SEO fields when you regenerate an existing product, and revert writes that pre-regeneration snapshot back to the live product. The page does not say every CSV row received a snapshot, and it does not state how long a snapshot is kept. Before you publish a bulk file, export the current descriptions if you may need them back. For products you regenerate one at a time, confirm the original is visible in the compare view before you push a field.
Questions about skipping one-by-one edits
Is there a 50-product cap per job?
Yes, on Pro. Basic is 10 products per job. Free has no CSV job. There is no higher tier in this catalog, and the cap is products, not words.
Can I revert a bulk publish?
Only where a pre-regeneration snapshot exists for that product. The catalog does not promise that for every CSV row. Keep your own export of the current descriptions before a large publish or import.
Should a large catalog leave Sidekick behind?
No. Keep Sidekick for the handful of products that are easier to draft in the admin. Use the CSV for the rows that are thin and numerous.
Batch the thin catalog on Basic or Pro, review the library, and keep Sidekick for the products that are faster one by one. Job sizes are listed on Squadexa AI.