AI Services Market › Use cases › Shopping & Errands
Shopping & ErrandsDone with custom-built or unnamed agent Oct 26, 2026
Order lunch on Uber Eats by tapping a picture
The author had an AI agent build a "Lunch" app where tapping a food photo spawns a subagent that orders the item on Uber Eats, to reduce decision fatigue. · Oct 26, 2026
- What They Did
- Colin asked his AI agent to build an app called Lunch. It shows a few pictures of foods he might want. When he taps one, the app spawns a subagent that places the order for that item on Uber Eats. The post doesn't describe the build steps or any manual intervention.
- How It Went
- Food arrives 60-90 minutes after the tap. The finished app is about 600 thousand lines of JavaScript, which is a lot of code for a simple ordering tool.
- Worth Knowing
- Expect a wait of 60-90 minutes between tapping and eating. The post is brief and doesn't give cost, setup details, or how the Uber Eats access works.
Try it yourself with custom-built or unnamed agent
Build me a simple personal 'Lunch' web app that shows photos of my [number, e.g. 6] usual lunch orders from [restaurant names or favorite dishes]. When I tap a photo, it should start a helper agent that opens Uber Eats in a browser and puts that item in my cart for delivery to [my delivery address]. The job is done when tapping a photo reliably gets the right item to the checkout screen; pause and show me the cart, total, and address before placing any order or paying.
Do this with a skill.
These AI Services Market playbooks do this exact job.