RoboTap – Automated Shift Finder

Raspberry Pi + OpenCV-guided robot with a custom 3D‑printed frame that detects and taps Amazon Flex shifts.

A Raspberry Pi‑powered, OpenCV-guided robot with a custom 3D‑printed frame that automatically detects and taps shifts in the Amazon Flex app. Designed and fabricated mechanical components in Fusion 360 while integrating embedded electronics, power management, and motion control. In a week of testing, shift-search time dropped from 15 hours to 1.95 hours.

RoboTap is a custom automation system designed to monitor and claim shifts within the Amazon Flex app. I built it to solve a personal workflow problem—manually checking for available shifts was consuming hours of my day, so I developed a robotic solution to handle the process autonomously.

What It Does

The device uses a Raspberry Pi 5 with a camera and OpenCV to detect new shifts on the screen. Once identified, a precision stepper-driven rail and electromagnet stylus perform the tap action through real-world motion control. The electronics include buck converters, relays, current protection, and servo control, all coordinated through custom Python code running on the Pi.

How It Works

I designed and fabricated the mechanical structure myself using a 3D printer. The process gave me extensive experience with CAD modeling in Fusion 360, particularly in designing modular mounts for motors, servos, and sensors while balancing strength, precision, and compactness. The project brought together embedded systems, mechanical design, and automation into a fully integrated prototype that eliminates repetitive manual work through intelligent hardware.

Why I Built It

Manually checking for available shifts was consuming hours of my day. I wanted an autonomous solution that could monitor the app and claim shifts as soon as they appeared, without me having to stare at the screen.

Highlights

  • Raspberry Pi 5 + camera + OpenCV for shift detection
  • Stepper-driven rail and electromagnet stylus for precise taps
  • Testing cut shift-search time from 15 hours to 1.95 hours in one week
  • Custom Python control, buck converters, relays, servo control
  • 3D‑printed mechanical design in Fusion 360
  • Modular mounts for motors, servos, and sensors