Shipping a Seasonal AI Side Project on Cloudflare Workers: Lessons from a Pet Halloween Photo App
Most side-project advice assumes you have time: launch, iterate, grow slowly. A Halloween product gets about three weeks. If people can't find it, trust it and pay for it in that window, the next chance is a year away. This October I shipped HallowPaws . You upload a photo of your dog or cat, pick a costume, and get an HD Halloween portrait that still looks like your pet. Below are the…
A Halloween side project called HallowPaws was launched, allowing users to upload a photo of their pet, choose a costume, and receive an HD Halloween portrait that still resembles their pet. The engineering and product decisions made during this short seasonal window were crucial for the success of the project. The stack used was TanStack Start + React 19, deployed to Cloudflare Workers D1 for data, R2 for images, a Cloudflare Queue for generation jobs, better-auth for accounts, and Stripe Checkout for one-time packs.
One of the main considerations was to keep the product simple for pet parents who are not prompt engineers. Therefore, there were no free-text input areas, and only 16 costume presets were available. Each preset had two creative fields, and the images were generated using a fixed template. This made the results reproducible and allowed for easy updates when a costume underperformed.
Another important decision was to watermark the images on the server before storing them. This prevented the model from treating the watermark as a suggestion and avoided any URLs for the previews, ensuring they were unique and not easily guessed. The image editing calls were slower than the Cloudflare proxy timeout, so generation was run as a job to avoid timeouts. The client would then poll the job status until it was completed.
To attract users during the short seasonal window, HallowPaws offered one free watermarked preview without requiring an account. Users could then sign up for one free HD credit, allowing them to unlock a clean, full-resolution portrait before making any payments. The pricing structure was designed to be fair for a product used for only a few weeks a year, offering one-time packs at various price points.
Lastly, the landing page featured honest before/after examples created using royalty-free pet photos that had been processed using the same presets as the customers. This ensured that the demos accurately represented the product's capabilities and maintained transparency with potential users.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
