Leaf AI is an independent, open-source project. If you find it useful, you can support its continued development at Leaf AI Support.
The Leaf Node is an open-source environmental monitoring and control device built on the ESP32-S3. It reads sensor data, publishes it over MQTT, and can trigger actuators — allowing you to monitor conditions and directly control connected devices such as lights, fans, pumps and valves. Configuration is handled wirelessly via BLE, making it a self-contained building block for precision agriculture and environmental monitoring and automation systems.
Leaf Node is built around one principle: it should just work. From first flash to live sensor data in minutes — simple enough for a grower to deploy in the field, open enough for a developer to extend and build on. No unnecessary complexity, no steep learning curve.
- Multi-protocol sensor support — I2C, RS485, SDI-12, OneWire on a single board
- MQTT telemetry — publishes sensor data and heartbeats, receives commands
- OTA firmware updates — triggered via MQTT command, no physical access needed
- BLE provisioning — initial Wi-Fi and MQTT setup from the companion app, no serial cable required
- Scheduling — DAC and actuator outputs can be programmed on a time-based schedule
- Smart control — MOSFET and relay outputs with duration-based switching
- Node chaining — UART-based daisy-chain for multi-node deployments
- Watchdog & health monitoring — automatic recovery with exponential backoff on sensor errors
- Factory mode — serial-console provisioning for production flashing
| Component | Details |
|---|---|
| MCU | ESP32-S3 |
| Status LED | WS2812B addressable RGB (v3) |
| RS485 | Half-duplex, DE/RE controlled |
| I2C | SDA GPIO8, SCL GPIO9 |
| SDI-12 | GPIO12 (shared with OneWire via hardware bridge) |
| OneWire | GPIO12 |
| Actuators | MOSFET (GPIO36), Relay (GPIO38), DAC (MCP4725 via I2C) |
| PWM | GPIO2 (1 kHz, 10-bit) + MOSFET gate (5 kHz, 10-bit) |
| Sensor | Interface | Wiring | Measurement |
|---|---|---|---|
| SHT31 | I2C (0x44) | +3.3V · GND · SDA · SCL | Temperature, Humidity |
| EZO pH | I2C (0x63) | +3.3V · GND · SDA · SCL | pH |
| EZO EC | I2C (0x64) | +3.3V · GND · SDA · SCL | Electrical Conductivity |
| TEROS 12 | SDI-12 | +3.3V · GND · SDI-12 | Soil moisture, temperature, EC |
| DS18B20 | OneWire | +3.3V · GND · Need Modification | Temperature |
| SLT5007 | RS485 | +3.3V · GND · A · B | VWC, Bulk EC, Soil temperature, Pore EC |
| CWT-PSS | RS485 | +12V · GND · A · B | PAR (Photosynthetically Active Radiation) |
| LEAFTHSN | RS485 | +12V · GND · A · B | Leaf wetness, Leaf temperature |
| CWT-SoilTHS | RS485 | +12V · GND · A · B | VWC, Soil temperature |
| CWT-THXXS | RS485 | +12V · GND · A · B | Air temperature, Humidity, VPD |
| Actuator | GPIO / Interface | Wiring | Typical Use |
|---|---|---|---|
| MOSFET | GPIO36 | +12V · -PWR | Pumps, fans, DC loads |
| Relay | GPIO38 | COM · NO · NC | AC/DC switching, high-current loads |
| DAC (MCP4725) | I2C (0x60) | 0-10V · GND | Analog 0–5 V output (e.g. dosing pumps) |
| PWM | GPIO2 | IO2 · GND | Variable speed / dimming control |
- KiCad project: hardware/pcb/kiCad/
- Enclosure STL/STEP files: hardware/enclosure/
No toolchain required. Works directly in Chrome/Edge.
- Get the hardware — order a ready-to-use Leaf Node at Build Your Leaf Node, or fabricate your own PCB from the KiCad files
- Open the Leaf AI Firmware Flasher and follow the on-screen steps
- Open the Leaf AI app, tap Add Device, and provision Wi-Fi via BLE
That's it — the node connects to the Leaf AI backend and starts publishing sensor data.
A free Leaf AI account is required to use the app. Sign up directly in the Leaf AI app.
For developers who want to modify or extend the firmware.
Prerequisites: PlatformIO (CLI or VS Code extension)
Fabricate the PCB from the files in hardware/pcb/ and wire your sensors according to the pin table above.
First flash: Use the Leaf AI Firmware Flasher for the initial flash. It writes a compatible serial number to the device that is required for communication with the Leaf AI app. Subsequent builds can be uploaded via PlatformIO.
git clone https://github.com/leafai-io/leaf-node-v3.1.git
cd leaf-node-v3.1
pio run --target uploadMonitor serial output (115200 baud):
pio device monitorOpen the Leaf AI app, tap Add Device, and follow the on-screen steps to configure Wi-Fi and link your account.
Once your Leaf Node is provisioned, open the Leaf AI app to view live sensor readings, configure schedules, and control actuators — all from your phone.
This project uses two open-source licences depending on what you are working with:
| Component | Licence |
|---|---|
| Firmware / Software | GNU AGPL-3.0 |
| Hardware (PCB, enclosure) | CERN-OHL-S v2 |
Software (AGPL-3.0): You can use, modify, and distribute the firmware freely, but any modified version you run as a networked service must also be released under the same licence.
Hardware (CERN-OHL-S v2): You can study, modify, and manufacture the hardware freely, but any modified design you distribute or produce must be released under the same licence with complete source files.
If you create a fork, we'd love to hear about it. Open an issue or send us a message so we can showcase your project and potentially integrate useful improvements.
Leaf AI is an independent, open-source project. If you find it useful, you can support its continued development at Leaf AI Support.