Summary
Chris Raroque shares a detailed update on Amy, his AI-powered calorie tracking app, which has crossed $3,000 in monthly recurring revenue at an 85% profit margin — unusually high for an AI app, where margins of 20–30% are more typical. The video focuses less on the revenue milestone and more on the technical and strategic work done to future-proof the app ahead of Apple’s iOS 27 Siri overhaul.
The centerpiece is a Siri integration built almost entirely through Claude Code in a single prompt session: users say “Hey Siri, log food in Amy,” speak their meal, and receive a nutrition preview before confirming. Raroque explains why Amy’s text-first input architecture made this integration far simpler than competitors like MyFitnessPal or Calai, which rely on search dropdowns or camera inputs that Siri cannot replicate. He also walks through building an MCP (Model Context Protocol) server for Amy, enabling AI agents like Claude and Codex to directly log food, analyze eating patterns, and make dietary suggestions — removing the current friction of exporting data and pasting it into a chat interface.
A secondary thread covers using OpenAI Codex for personal meal planning, including having it autonomously order groceries from Whole Foods (with a cautionary note about a 4-pound lemon order). The video is a practical snapshot of how a solo developer is using the current AI toolchain — Claude Code, Codex, and MCP — to build and maintain a competitive consumer app.
📺 Source: Chris Raroque · Published June 30, 2026
🏷️ Format: Workflow Case Study
