Alpha · Paused
Where should I get dinner tonight?
Local pickup, compared
SavageJunk started with a practical question: where can I get the best-value burger, pizza or takeaway nearby? I wanted to compare restaurants without opening several apps and websites. The focus is local pickup around Haugesund, using the restaurants’ own prices whenever possible.
The WordPress web application has a custom backend for importing and normalising menu data. It includes burger, pizza, deal and favourite views, with sorting by price, distance and value. Branches have their own prices, while favourites follow the product across a chain.
Different sources, different solutions
The features include Deal Score, price history and price-drop detection, product images, restaurant logos and links to the restaurant’s own ordering system. An Android version was also started, with the web interface inside a native app.
Bankers Pizza needed browser capture of its rendered menus and images. McDonald’s combined its public national catalogue with separately captured branch prices. For Burger King, I identified the first-party API used by the Norwegian Android app to retrieve local menus, prices and product data.
The difficult part was the data
Every chain exposes its menus differently. Dynamic websites, changing structures and temporary image URLs made the importers fragile. Some server-side experiments overloaded the hosting account and caused 503 errors; image failures also exposed frontend problems.
Many issues were resolved, but the alpha still has missing images and imperfect records. The central problem remains the manual work needed to discover and maintain each source. The screenshot shows the actual application, including those limitations.
Paused, with a different next step
The project is paused rather than abandoned. Before adding more chains, I want to explore agents that can inspect changed sources and adapt data collection with less manual work. That is a direction for the next version, not an autonomous system already running.