Adaptive AI Ambient Light With ESP-Claw + UNIHIKER K10 (Zero Code)

This project builds an adaptive, AI-controlled ambient light using the ESP-Claw framework on a UNIHIKER K10 with a Gravity SCI data-acquisition module, with zero coding and low hardware cost.

By connecting a light sensor and an RGB strip, the system automatically adjusts brightness and on/off state from real-time ambient light: on when the room gets dark, off when it is bright enough, brighter in very dim conditions. ESP-Claw's LLM understands the environment and decides the lighting, so it behaves like a real AI ambient-lighting agent rather than a fixed-threshold automation.

1. UNIHIKER K10 x 1
2. Gravity: SCI Data Acquisition Module x 1
3. Ambient Light Sensor x 1
4. 7-LED RGB light strip x 1
5. 4-pin I2C cable and jumper wires

Connect the SCI module to the I2C port of the UNIHIKER K10 with a 4-pin cable. Attach the light sensor to Port2 of the SCI module, and plug the RGB light strip into the P1 port of the UNIHIKER K10.

Flash the ESP-Claw firmware to the UNIHIKER K10 following the companion guide, How boring the sensor-free ESP-Claw is, then complete the Wi-Fi and LLM setup. Once configured, you can enable ESP-Claw to "perceive ambient light".

First, make ESP-Claw understand the live light data. Send this message in your chat tool:

I have connected a light sensor to the SCI module. Please read the current light intensity and tell me whether the room is dark, dim, normal, or bright.

ESP-Claw reads the sensor through the SCI module, handles the unit conversion, and classifies the room from the real-time value. From this point on it is observing the physical environment, not just reading text.

Now let it control the light. Send:

I have connected a 7-LED RGB strip to port P1 (GPIO2) of the UNIHIKER K10. Please create an adaptive lighting strategy that adjusts brightness to the ambient light and switches colors by time of day.

ESP-Claw parses the request, generates the GPIO control logic, and runs the strategy when conditions are met. Unlike a fixed "turn on below X lux" rule, its LLM can weigh time, weather and habits to decide whether light is actually needed and at what brightness, which is the core difference between an AI agent and conventional IoT automation.

The light sensor detects changes, the SCI module standardizes the data, ESP-Claw interprets the environment and decides, and the UNIHIKER K10 drives the RGB strip: a full loop from perception to action. Once ESP-Claw can "understand light", the lamp is no longer a simple on/off device. It becomes an AI lighting agent that senses its surroundings and responds on its own.

Want to go further? Check out the companion builds: an AI plant-care assistant and the sensors and SCI module guide.