SCARAScribe is a two-axis SCARA robotic arm designed to reproduce user-drawn strokes onto paper with guaranteed single-stroke accuracy, validated through real-time computer vision feedback. The system enables a user to draw a continuous stroke on a web-based canvas, which the arm then physically reproduces on a 7.25" square paper workspace while continuously monitoring execution quality. The system integrates an ESP32 microcontroller driving TMC2209 stepper motor drivers, a Python FastAPI backend with OpenCV-based vision processing, and a browser-based GUI connected over WebSocket. ArUco marker calibration provides sub-millimeter workspace registration, while a multi-source fault detection pipeline (vision-based paper boundary monitoring, cross-track deviation analysis, StallGuard motor stall detection, and hardware limit switches) ensures the arm halts immediately if drawing accuracy is compromised. Key accomplishments include the implementation of a deterministic 8-state machine with a universal recovery hub, a modular three-layer architecture enabling independent testing of vision, motion, and firmware subsystems, and an accuracy reporting system targeting ER-2 compliance (< 0.25" deviation on 10" lines). The primary challenge was hardware and our choice not to make a custom PCB, this led to some major pitfalls down the line, and we plan on implementing a v2 that addresses this (along with z-axis lift).