[ESP-Claw and UNIHIKER K10]No Need to Buy a Smart Flower Pot: Build Your Own AI Plant Care Assistant

Project Introduction

Smart plant care products such as self-watering pots and plant monitoring systems are becoming increasingly popular. However, most of them are expensive and only provide fixed functions with limited customization.

With ESP-Claw, UNIHIKER K10, and the SCI Data Collection Module, you can now build your own AI Plant Care Assistant at very low cost and with zero coding.

In this project, ESP-Claw uses a soil moisture sensor to monitor plant conditions in real time, while autonomously handling watering control, environmental analysis, plant status logging, and IoT data uploads through natural language interaction.Instead of writing complex code or automation rules, you simply tell the AI Agent:“Automatically water the plants when the soil becomes too dry, and generate a daily plant report.”

The rest — sensing, analysis, control, and long-term monitoring — is handled automatically by ESP-Claw and the UNIHIKER K10.For anyone interested in AI Agents, smart hardware, or AIoT, this is a simple and low-cost way to build an AI system that can actually “take care of plants.”

Hardware Connection

Connect the SCI Data Collection Module to the I2C interface of the UNIHIKER K10 using a 4-pin cable, then connect the soil moisture sensor to Port1 of the SCI module.

Next, use a 3-pin cable to connect the relay module to the P0 interface of the UNIHIKER K10. Connect the relay module’s VIN terminal to the power supply and connect the VOUT terminal to the DC water pump.

Finally, set the relay switch mode to: NO (Normally Open)

Project Workflow

After completing the hardware wiring, follow the tutorial ‘How boring the sensor‑free ESP‑Claw is’ to flash the ESP-Claw firmware onto the UNIHIKER K10.

According to the official SCI Module Wiki, configure the soil moisture sensor as an analog input sensor. The SCI module is not only responsible for analog data acquisition, but also acts as a bridge between the physical world and the AI Agent. It automatically standardizes sensor data, records timestamps, and converts raw sensor readings into structured environmental information that can be understood by the AI Agent.

Next, we will gradually transform the ESP-Claw-powered UNIHIKER K10 from a passive device into a real Plant Care AI Agent. The entire process is divided into four stages.

Step 1: Soil Moisture Monitoring for Plants

First, instead of immediately implementing automatic control, we let the Air Board K10 “understand the data” first.

Send the following prompt to the Air Board K10 via a chat tool:
“I have connected a soil moisture sensor to the SCI module. The raw reading ranges from about 0–2000 mV. Please map it into soil moisture values so that I can read the soil moisture level in my garden.”

Through this step, the raw sensor output is converted into soil moisture data associated with plant conditions, enabling the Air Board K10 AI Agent to build an understanding of plant states.

Besides reading soil moisture, the AI Agent can also analyze environmental data over a period of time and generate plant‑growth environment reports based on timestamped standardized data from the SCI module.

Send the following message to the UNIHIKER K10 via the chat tool:

Collect one minute of soil‑moisture data, display the line graph on the UNIHIKER K10 screen, and generate a plant‑growth environment analysis report.

You will see the AI Agent on the UNIHIKER K10 respond accordingly.

Step 2: Establish Automatic Water Pump Control Rules

Based on environmental perception, we will now connect actuators to the UNIHIKER K10 AI Agent, enabling it to independently control the water pump according to soil moisture.

Send the following message to the UNIHIKER K10 via the chat tool:

The water pump is now connected to the P0 port (GPIO1) of the UNIHIKER K10. Please create an automation rule for me: Monitor the soil moisture every 15 minutes. When the moisture drops below 25%, automatically turn on the water pump to water the plants. Actively send monitoring information and execution result records after each operation.

In this way, you can use natural language to control the UNIHIKER K10 AI Agent to perform automatic watering tasks on a regular basis.

Step 3: Automatic Data Upload to IoT Platform

So far, ESP-Claw has realized soil moisture detection and automatic watering control. You can upload collected soil moisture data and pump operation records to the IoT platform. This allows remote data viewing. Meanwhile, the AI Agent on UNIHIKER K10 can track real-time data changes and gain long-term data storage capabilities, so as to analyze moisture trends, watering frequency and plant water consumption patterns.

I asked my UNIHIKER K10 AI Agent for recommendations on suitable IoT platforms, and it suggested several options. For this project, I chose ThingSpeak

Following the setup guide, I first created a ThingSpeak account and then created a new channel for the plant monitoring system.

Click “New Channel” and configure the channel as follows:

Channel Name: ESP-Claw Garden Monitor

Field 1: Soil Moisture (%)

Field 2: Raw Voltage (mV)

Field 3: Moisture Level

Field 4: Pump Status

After completing the configuration, click “Save Channel”. ThingSpeak will automatically generate the corresponding API Keys for this channel, which will later be used by ESP-Claw to upload sensor data and watering records to the IoT platform.

Click “API Keys” to view the channel’s API information. Copy the Write API Key.

And send it to the ESP-Claw AI Agent, allowing the UNIHIKER K10 to automatically upload soil moisture data and watering records to the ThingSpeak IoT platform.

After the configuration is completed, you can see that the corresponding data fields have been created automatically on the IoT platform, and real-time soil moisture data and watering records sent from the UNIHIKER K10 are now being received successfully.

Step 4: Install the Plant Care Assistant Skill

At this stage, the UNIHIKER K10 AI Agent has evolved into a system capable of environmental sensing, autonomous action execution, long-term data logging, and trend analysis. We can now package all these capabilities into a complete automated plant monitoring system.

Send the following instruction to the UNIHIKER K10 through the chat interface:

Please compile all the above functions into a complete Plant Care Assistant Skill and install it automatically. The system should perform soil moisture monitoring every 15 minutes, automatic water pump control, IoT data uploads, and automatic plant monitoring report generation.

After the installation is completed, the ESP-Claw-powered UNIHIKER K10 will continuously monitor plant conditions, autonomously control watering behavior, upload environmental data to the cloud, and actively generate long-term plant care reports, turning it into a real AI Plant Care Agent capable of interacting with the physical world.

Conclusion

In this project, we used the ESP-Claw AI Agent framework and the UNIHIKER K10 to build a Plant Care AI Agent that evolves beyond traditional rule-based automation.

The soil moisture sensor continuously monitors the real-world environment, while the SCI Data Collection Module standardizes and structures raw sensor data into environmental information that can be understood directly by the AI Agent. Powered by ESP-Claw, the UNIHIKER K10 is responsible for environmental understanding, decision-making, and autonomous action execution.

Once ESP-Claw gives the UNIHIKER K10 the ability to “understand” the physical world, the hardware is no longer just a device that executes commands — it becomes a real AI Agent capable of taking care of plants autonomously.