cshitian/antigravity-task-led

Antigravity Task LED is an IoT project that turns the activity state of an AI coding assistant into physical light signals on your desk. When the AI in Antigravity IDE or VS Code is generating code, finishes successfully, or hits an error, a small circuit with two LEDs, one red and one green, reacts with different lighting patterns in real time. The system works in three stages. An editor plugin monitors the AI's current state. When the state changes, it sends a lightweight HTTP request to a Chinese cloud IoT platform called Bemfa. Bemfa forwards the message over MQTT to an ESP8266 microcontroller, which is a small, inexpensive WiFi-capable chip that drives the two LEDs directly. The project includes 19 different LED patterns, each with its own character. A few examples: alternating breathing, where the two LEDs fade in and out opposite each other, a Tai-Chi S-curve pattern using a third-power sine wave for an unusually smooth crossfade, an ECG simulation that mirrors the shape of a real electrocardiogram, an SOS Morse code sequence, and a radar scan pattern where a slow green sweep transitions into a rapid red lock signal. All patterns are implemented without blocking delays, using a milliseconds-based state machine so the microcontroller stays responsive to new MQTT messages at all times. The hardware requires an ESP8266 development board (such as a NodeMCU), two LEDs with 220-ohm resistors, and a Bemfa account for the cloud relay. Credentials are stored in a separate secrets file excluded from the repository. Setup involves flashing the Arduino firmware to the board, then installing the editor plugin as a .vsix package in your IDE. The project includes complete source code for both the Arduino firmware (580 lines) and the TypeScript editor extension. The README is written in both Chinese and English.