DEV Community

Cover image for i run two Shopify stores with zero manual product imports. here's the stack.
lofder.issac
lofder.issac

Posted on • Originally published at github.com

i run two Shopify stores with zero manual product imports. here's the stack.

#shopify #ecommerce #automation #ai

I've been running two Shopify stores for about a year now. One does pet accessories, the other home gadgets. Between them I move maybe 60 products sourced from AliExpress and Alibaba.

The first three months were brutal. Finding products, importing them, editing titles, setting prices, pushing to Shopify — all by hand. I'd spend entire evenings copy-pasting AliExpress URLs, fixing titles like [2026 HOT!!!] Premium Quality Pet Grooming Glove FOR CAT DOG Hair Remover, and manually calculating markups.

Then I found a stack that cut all of that to near-zero manual work.

DSers handles the product pipeline

If you're sourcing from AliExpress and selling on Shopify, DSers is the middleman between your supplier and your store. It connects your Shopify (or Wix) account, lets you import products from AliExpress / Alibaba / 1688, manage variants, and bulk-push to your store.

I tried other tools before landing on DSers. Oberlo was sunset. CJ Dropshipping has its own supply chain which is nice but limiting. DSers just felt right — fast UI, solid AliExpress integration, and the pricing rule system is underrated. You set a formula once (2.5× supplier cost + $2 buffer, compare-at at 3× for that "sale" look) and it applies across every import.

The free tier covers everything I need for two stores.

The bottleneck was always me

DSers solves the infrastructure — but someone still has to sit there and do the work:

  • Find products worth selling
  • Import them one by one (or in batches, but you still need the URLs)
  • Clean up titles and descriptions for SEO
  • Set pricing rules for each import
  • Push to the right stores with the right visibility settings
  • When a supplier raises prices or goes out of stock, find a replacement, compare every SKU variant, and update the mapping

That last one was the real time sink. I wrote about the supplier replacement problem in a separate post — the short version is that matching variant options across suppliers (where one calls it "EU PLUG (220-240V)" and another says "European Standard") is error-prone and tedious.

AI agent layer: MCP server for DSers

This is where the second piece comes in. I built an open-source MCP server that wraps the entire DSers workflow into tools that any AI assistant can call.

MCP (Model Context Protocol) is Anthropic's standard for connecting AI to external tools. If you've used ChatGPT plugins or function calling, same concept but open and standardized. Claude, Cursor, Windsurf, Cline — they all support it.

With the server running, I talk to my AI assistant in plain English:

Import this product: [AliExpress URL]. Apply 2.5× pricing, clean up the title, push to both stores as draft.

One sentence. It calls the DSers API behind the scenes, applies my rules, and the product shows up in my Shopify admin ready for review.

For supplier replacements:

Check all products in store st-102. Anything with a margin below 20%, find a cheaper supplier and show me the comparison.

It searches, scores candidates across five dimensions (spec overlap, option coverage, title similarity, image similarity, market signals), matches every SKU variant, and gives me a report. If the match confidence is high enough, it can auto-update the mapping. If not, it asks me to decide.

The actual workflow now

Monday mornings I spend about 20 minutes with my AI assistant:

  1. "Any supplier price changes this week?" — it checks all products, flags anything that moved
  2. "Import these 5 products" — paste URLs I collected during the week from my research
  3. "Push everything that's been sitting in staging for more than 3 days" — clears the import queue

That's it. The rest of the week I focus on marketing and customer support instead of product management.

Setup if you want to try it

DSers — sign up here, connect your Shopify store. Free tier works fine.

MCP server — add this to your Claude Desktop config (or Cursor, Windsurf, whatever you use):

{
  "mcpServers": {
    "dsers": {
      "command": "npx",
      "args": ["-y", "@lofder/dsers-mcp-product"]
    }
  }
}
Enter fullscreen mode Exit fullscreen mode

First time you run it, the server opens a browser tab for DSers OAuth login. Your credentials never touch the tool.

13 tools total: product search, single and batch import, variant editing, pricing rules, multi-store push, safety checks, and the automated supplier replacement I mentioned. Full docs on GitHub.

Honestly the biggest win isn't the time saved — it's the consistency. I don't forget to check supplier prices anymore. I don't rush through variant matching and accidentally ship wrong items. The AI doesn't get tired at 11pm and mix up a US plug with a UK plug.

If you're running a similar operation and still doing imports by hand, the combo of DSers + this MCP server might save you a few evenings a week. And if you're already using DSers but haven't tried connecting an AI agent to it — that's the part that made the biggest difference for me.

What does your dropshipping stack look like? Always curious how others handle the supplier management side.

Top comments (0)