I automate ad creative production for online sellers.
Dropshipping lives on testing many creatives quickly. Making each one by hand is slow and expensive. I built a workflow that does the heavy part automatically.
The problem
Sellers need a constant stream of fresh video and image ads for each product. Shooting, editing and revising them takes hours per product, or costs money per creative.
What I built
An automated workflow. Photos of a product from several angles go in. The workflow submits them to Higgsfield's API, which runs the Seedance model and returns finished video and image ads.
What it gives you
Ready-to-test creatives from the photos you already have, so you can try more products and more angles with far less manual work.
Four steps, no editing software.
The whole flow runs automatically once the photos are uploaded.
Upload photos
Add several photos of the product taken from different angles.
Automation starts
The workflow collects the photos and builds the generation request.
Generation
Higgsfield's API runs Seedance, which creates the video and image ads from the photos.
Download
The finished creatives are collected and ready to upload to your ad accounts.
From product photo to finished ad.
Drag the slider to compare the original photo with the generated ad.
Built on Higgsfield's API and Seedance.
I don't train models. I connect strong existing ones into a workflow a seller can actually use.
Higgsfield API
A generative media platform with an API. My workflow uses it to submit generation jobs programmatically and receive the finished video and image files.
Seedance
ByteDance's video generation model, accessed through Higgsfield. It takes reference images and instructions and generates short, ad-style clips.
Workflow automation
The glue that matters: it takes the uploaded photos, runs the generation, and delivers the results without manual steps in between.
Measured, not promised.
Numbers from real runs of the workflow.
What shipped, and when.
I build in the open. Each entry is dated.
Built by one person, in Istanbul.
I'm Muhammed, a computer programming student focused on AI automation for small online businesses. This project started from a practical problem in dropshipping: good ad creatives are slow to produce, and sellers need many of them.
I'm developing it step by step and publishing real examples and numbers as they come in.