AI Currency Detection Pipeline for Visually Impaired Users leverages YOLOv8 deep learning on OCI-integrated Raspberry Pi 5 edge devices to detect and vocalize Indian currency denominations in real-time. Raw JPEG frames from a CSI-connected camera module flow through OpenCV preprocessing, tensor normalization, and YOLOv8 object detection, achieving ~40ms latency for INR 10–500 denomination classification. The pipeline outputs UTF-8 text converted to offline speech via pyttsx3, delivering immediate audio feedback through connected speakers. This architecture demonstrates how OCI can orchestrate edge ML inference for accessibility, combining hardware GPIO inputs, on-device model serving, and low-latency audio synthesis. Fork and customize this diagram on Diagrams.so to adapt the pipeline for other currency systems, integrate OCI Functions for cloud fallback, or extend with additional sensor inputs.