Skip to content
Planti

Plant care with a memory

Less guessing.
More growing.

Planti is a plant-care agent that learns from your plant's history. It combines live sensor readings with trusted horticultural research, tells you what your plant needs, and checks whether the care actually worked.

planti live

Soil moisture

42%

23.4°C
drying slowly

Sent 12s ago · next reading in 48s

How it works

A care loop that closes itself

Most plant apps stop at a reminder. Planti follows through: it watches what happens after you act and keeps that in your plant's memory for next time.

  1. 01

    Observe

    A soil sensor reports moisture and temperature every minute, and each reading lands in your plant's history.

  2. 02

    Analyze

    Planti calculates trends, drying rate and time since watering with plain, testable code. No guessing at the numbers.

  3. 03

    Recommend

    NVIDIA Nemotron reads that evidence and explains what your plant needs, with a confidence level you can see.

  4. 04

    Act

    Log what you did: watering, feeding, moving it to more light. Care events sit right on the moisture chart.

  5. 05

    Verify

    Planti checks whether the care worked by comparing the readings before and after, then remembers the outcome.

Plant Memory

Every reading, recommendation and care event becomes part of a timeline written in plain language. Planti learns how fast your pot dries out, how your plant responds to watering, and what worked last time, so advice gets more specific the longer you use it.

Honest about every number

Every value on screen is labeled by where it came from, so you always know what is a fact and what is a judgment.

Measured
Straight from the sensor in your pot.
Simulated
Demo data, always marked as simulated.
Calculated
Deterministic math on your readings.
Interpreted
The model's reading of that evidence.

The hardware

A sensor in the pot, a screen on the shelf

Planti runs on a LILYGO T-Display S3 with a capacitive soil sensor. It shows moisture and temperature on its own screen and uploads a reading every minute over Wi-Fi. Pair it to a plant in a few clicks. No sensor yet? The dashboard works with simulated data so you can try the full loop first.

  • ESP32-S3 board with a built-in display
  • Capacitive soil moisture and temperature sensor
  • Per-device credential, readings sent over HTTPS

Give your plant a memory.

Your plants, readings and care history stay private to your account.

Get started
Planti · Built for the Nebius × NVIDIA Global AI HackathonLess guessing. More growing.