How I uploaded more than 500 products to PrestaShop in one day with Claude
Here's exactly how it went. A client handed me a CSV with more than 500 products to set up their PrestaShop store: photos, descriptions, prices and metadata, all to be loaded. The "normal" route was weeks of manual work or a development quote. I solved it in one day, directing Claude, without a professional dev team and without writing the code by hand.
The challenge: 500 products in a CSV and an empty PrestaShop
It's the classic starting point when an ecommerce launches or migrates: a supplier or client CSV with hundreds of references and a store waiting to be filled. Each product needs a full listing: name, short and long description, price, category, images and the metadata (meta title and meta description) so Google understands it.
Multiply that by 500 and you have the real problem: by hand it's weeks, and a copy-paste mistake repeats itself five hundred times.
Why I didn't load it by hand (or pay for weeks of development)
Uploading 500 listings by hand isn't work, it's punishment. And building custom development meant time and cost before seeing the first product live. I wanted the catalog working now, with a solid SEO base from product one and no descriptions copied from the supplier (that's duplicate content and Google penalizes it).
PrestaShop does have a native CSV import (Advanced Parameters → Import). You select the entity to import (products), upload the CSV file, set the field separator (a semicolon by default) and the multiple value separator (a comma), map each column to its field and check the data before importing. It's fine for clean data, but it doesn't write a single description, and with hundreds of rows one badly formatted column (prices with decimal commas, a wrong tax rule ID, image URLs that don't respond) is enough to stop the run.
The solution: connecting Claude with PrestaShop
Here's the interesting part. Instead of a classic development project, I set up a connection with PrestaShop through Claude that turns the converted CSV into complete product listings.
The system didn't just "upload rows." For each product it generated unique descriptions (not copied from the supplier), assigned a category, linked its photos and filled in metadata built for search. Not a catalog dumped any old way, but listings ready for SEO from the start.
My job wasn't writing code line by line: it was directing the process. Defining exactly what listing I wanted, how it had to be structured for SEO, reviewing what Claude proposed, fixing what didn't fit and running it again. When something broke, I passed it the error and it was solved in seconds.
The result: a full catalog in one day, with SEO baked in
In a single day, the 500+ products were live in PrestaShop with their photos, descriptions, categories and metadata. What in a traditional flow takes weeks (or a development invoice) was solved in hours. And the client didn't start with a catalog thrown together, but with listings ready to rank.
What I learned
- AI doesn't replace business knowledge, it multiplies it. It worked because we knew what listing we wanted and how it had to look for SEO.
- You don't need a professional dev team to automate a bulk upload: you just need to direct the process well.
- Let the machine do the repetitive part. The value is in the strategy, not in copy-pasting 500 listings.
If you run a PrestaShop store (or Shopify) and you're facing a bulk product upload, it's solved fast and well today. It's part of what we work on inside ecommerce SEO and Universo Ecommerce.
Frequently asked questions
Yes. With a well-built process, hundreds of listings with photos, descriptions and metadata load in hours instead of weeks. The key is preparing the CSV file well and defining how each listing should look.
No. What matters is having a clear goal, structuring the listings for SEO, directing the tool and reviewing the result. That direction is exactly what an agency brings.
They shouldn't. The process generates original descriptions; copying the supplier's creates duplicate content and Google penalizes it.
Four of them: the entity (Products), the field separator (a semicolon by default), the multiple value separator (a comma by default) and whether to use the product reference as the key, so a second import updates products instead of duplicating them. Then map each column to its PrestaShop field and check the preview. Test with ten rows first: if those import cleanly, you've ruled out most format errors before the big run.
Usually for one of three reasons: the file isn't saved in the encoding the importer expects (UTF-8) and accents break, a column holds data in the wrong format (decimal commas, tax rule IDs that don't exist, broken image URLs), or the file is too large for the server's execution time. Split the CSV into smaller batches, fix the encoding and validate the columns before importing. Downloading images is the slowest step, so batches with many photos need extra margin.