Overview
DogArt.com is a playful, powerful web platform that allows dog lovers to turn their pets into AI-generated works of art. The project involved integrating advanced AI image generation, credit-based billing, and a seamless user experience all delivered by Magnifica.
Background
Concept behind the project: “What if anyone could upload a few photos of their dog and receive stunning, AI-generated portraits in any style they imagined?”
Whilst ChatGPT and orther AI tools are great a generating images it’s not always clear how you can engineer a prompt to get the results you want. Also, not everyone is ready to use a generic AI app.
The vision was to create a product that took the complexity and confusion out of using AI to generate amazing dog pictures, that actually looked like the users own dog!
The Challenge
With the concept agreed it was time to look at what was required to make DogArt.com work. We needed more than a pretty website, we needed a full-stack digital product that could:
- Let users train an AI model on their dog’s photos
- Generate realistic, high-quality images in seconds
- Offer a simple ecommerce experience with credit-based billing
- Be secure, scalable, and maintainable from day one
After some initial research and AI proof-of-concept work to figure out how we could allow users to train a custom model, we had a solution designed and ready to build.
The solution
We worked quickly to shape a fast, frictionless product experience powered by cutting-edge AI. The component parts are as follows:
AI + API Integration
We integrated with Replicate’s API to manage AI model training and image generation. This allowed users to:
- Upload up to 10 reference photos
- Train a custom AI model for their dog
- Prompt new image generations with styles like “Van Gogh”, “Superhero”, or “Royal Portrait”
We wrapped these AI operations in a robust Azure Function backend to handle timeouts, retries, and webhook-based callbacks.
Credit System & Stripe Payments
Rather than charge per image, we designed a credit-based billing system:
- 10 credits to train a model
- 1 credit per image generation
Credits are purchased via Stripe Checkout, and we implemented webhook handling to:
- Apply credits on successful payments
- Track purchases and usage in a secure SQL database
- Send confirmation emails using SendGrid
UX & Frontend
Built with ASP.NET Core MVC and Tailwind CSS, the frontend is clean, fast, and optimised for mobile. We focused on:
- Clear onboarding and uploading process
- Progress indicators during model training
- Instant feedback when generating new artwork
- A gallery-style layout with downloadable image previews
Cloud Native Infrastructure
Deployed on Microsoft Azure, the platform uses:
- Azure Blob Storage for image handling
- SQL Server for transactional and user data
- Azure Functions for background tasks
- SendGrid for transactional email
This approach made DogArt.com it cost-efficient and scalable from day one.
The Result
DogArt.com launched in May 2025 and saw immediate engagement from dog lovers across the world. Here are some examples!
The next steps for the business are to launch a range of sample prompts to make generating great mages even easier. There are also plans to extend the shop area to allow customers to print their AI photos onto a range of items such as t-shirts, mugs, cushions etc.
Why It Worked
DogArt.com is a perfect example of how emerging technology, when paired with user-centred design and a smart business model, can deliver something delightful and commercially successful. Magnifica delivered:
- A lean MVP in weeks, not months
- Resilient backend architecture
- Fast, joyful frontend UX
- Strategic advice around AI implementation and monetisation
Why We Deleted dogart.com
With the advances in AI and the fast pace of change it became clear that whilst the offer was strong consumers could use any number of free or almost free AI services to create similar outputs. No problem, but we are about action at Magnifica, so we decided to delete dogart.com and move on with other more profitable projects.