PlantHub watches soil, light, temperature, and humidity in real time — then tells you, in plain English, what your plant needs and why. No more guesswork. No more crispy fiddle-leafs.
Soil moisture at 24%. I'll water for 8s in the morning — rain is expected tomorrow.
The Problem
You shouldn't need an engineering degree to keep your tomatoes alive — or $200 to automate one plant.
80%
of houseplants die from improper watering.
$200+
is the average cost of a smart garden system.
Hours
of YAML and soldering for any DIY solution.
Timers
is all most "smart" planters really are.
Meet PlantHub
Reads sensors and the forecast, then picks from five AI modes — Conservative through Manual. A 6-stage learning loop with semantic memory means every decision is informed by past outcomes, not just rules.
Premium hardware from $15. No lock-in, 5–10× cheaper than the competition. Safety limits on watering time and cooldowns are enforced in code, not prompts.
AI compiles schedules into on-device rules. Wifi drops, your garden keeps going — and skips watering when rain is in the forecast.
Push notifications to Android, iOS, and web browsers via AWS SNS. Threshold warnings, AI decisions, device-offline events — with quiet hours and per-severity control.
How it works
Three steps from box to autonomous garden. No YAML. No soldering. No regret.
Click a button in your browser — Chrome, Edge, Brave, or Opera — and PlantHub installs the firmware in about 30 seconds. No drivers, no Arduino IDE, no soldering. The node self-registers and shows an 8-character claim code.
Flash now in your browserPaste the code, pick your plant. PlantHub loads its profile with AI-tuned ideal ranges, seeds starter automation rules, and the AI starts generating smarter ones from day one.
Five AI modes from fully autonomous to notify-only. Override per tenant, per zone, or per plant — anytime. The AI adapts its check schedule to your plant’s needs.
Your AI is currently learning about your Basil's care, showing promising results in watering with a typical growth response. Recent watering actions have positively impacted its health, indicating a foundation for refining its needs further.
Basil prefers consistently moist soil between 40-70%.
Optimal temperature for basil is between 20-30°C.
Maintain humidity levels around 50-70% for healthy basil.
Basil thrives in bright indirect light, ideally 500-1200 lux.
Live demo
This is what a real PlantHub device page looks like. Sensor data flows in via SSE, AI evaluates your plant's condition, and actions are recommended — all in real time.
Capabilities
Skips watering when rain is on the way. Pre-empts heatwaves. Tells you why before acting.
Ask "Why are my basil leaves drooping?" and get answers backed by your sensor history, weather data, and learned patterns — in plain English.
Analyzes your sensor data and plant profile, then recommends threshold-based automation rules you can approve with one tap. No manual config needed.
Scores each actuator's effectiveness, climbs from LOW to EXPERT confidence, and factors your feedback into the next run.
Compiled from 30 seeded templates and AI recommendations. Your wifi can drop all afternoon — the garden keeps going.
Conservative, Balanced, Aggressive, Power-Saver, Manual. Override at tenant, zone, or device — inheritance sorts the rest.
Checks more often when your plant is stressed, less when it's thriving. Adapts its own interval to save energy and API costs.
Decisions, threshold warnings, and device-offline events streamed over SSE — plus push notifications straight to your phone via SNS.
Moisture, temp, light, pH ranges for tomatoes, herbs, succulents and more. Create custom profiles with your own ideal ranges.
Tell PlantHub what you're growing and it recommends ideal min/max values for every sensor — calibrated to your specific plant species.
One-tap AI summary across all your devices. See which plants need attention, what's thriving, and what the AI plans to do next — in plain English.
AI learning loop
Most "smart" planters guess once and repeat forever. PlantHub watches what happened, remembers similar situations, discovers what works together, and tightens its own rules. Confidence progresses through four tiers: NO_DATA → LEARNING → ADVANCED → EXPERT.
Every action captures the sensor delta it produced — before vs. after, per actuator.
Effectiveness is scored. Confidence climbs LOW → MEDIUM → HIGH → EXPERT as samples accumulate.
Your feedback — "too early", "too long" — is batched hourly and turned into learned context.
Each evaluation is vectorized and stored. Before the next check, the AI recalls similar past situations to inform its decision.
Detects multi-actuator patterns — like watering + ventilation at 6 AM producing 2× better results than either alone.
All learned patterns — single, multi-actuator, and your preferences — are injected into the next prompt.
Safety you can prove
These aren't prompt instructions — they're enforced inside the code path that delivers every actuator command. The LLM physically can't exceed them.
Max water duration
Hard cap per command — AI cannot override.
Per-actuator cooldown
Same actuator can’t fire again until the timer expires.
Daily evaluation cap
Defaults to 4 per device per day. You set the limit — prevents runaway LLM spend and rapid decision churn.
Confidence ladder
Real thresholds, not vibes. Confidence combines four weighted factors: effectiveness, sample size, consistency, recency — and the AI's behaviour shifts at each tier.
< 5 samples
Just claimed. Uses the seeded plant profile and conservative defaults.
≥ 5 samples · ≥ 40% confidence
Detects patterns — typical watering durations, sensor response time, time-of-day effects.
≥ 20 samples · ≥ 70% confidence
Multiple weeks of data, tight estimates, consistent outcomes. Proposes rule changes.
≥ 50 samples · ≥ 85% confidence
Knows your plant in your environment. Surfaces cross-actuator patterns. LLM-free routine decisions.
Comparison
We love ESPHome — we built on it. We just made it accessible to people who aren't engineers.
| Feature | PlantHub | ESPHome | Home Assistant |
|---|---|---|---|
| Setup time | Flash & claim | Hours of YAML | Complex local setup |
| Intelligence | AI-based (context aware) | Rule-based | Rule-based |
| AI chat (ask about your plants) | |||
| AI-suggested automation rules | |||
| AI-tuned sensor ranges | |||
| AI memory (learns from history) | |||
| Semantic memory (recalls similar situations) | |||
| Plant knowledge | |||
| Weather integration | Manual add-on | ||
| Offline operation | Cloud-dependent | ||
| Push notifications | Native SNS (Android/iOS/Web) | You build it | Complex YAML |
| Self-provisioning hardware | Flash & claim | Flash firmware | Flash firmware |
| Mock device profiles | 3 built-in | ||
| Data retention | 1 year of history | Run your own DB | Your DB, your problem |
| Safety limits | Fail-safes (AI can’t override) | Only what you wrote | Only what you wrote |
| Mobile app | Native Android + Web | Clunky third-party | |
| Hardware included | |||
| Customer support | Forums |
Hardware
The same chips expensive competitors use — without the 10× markup or locked-down firmware.
Starter Node
1 sensor
Garden Kit
Soil + env + pump
Greenhouse
5 nodes
Learn more about ESP32 garden automation, explore our sensor node specs, or see how ESP32 + AI powers intelligent plant care. Hobbyists can grab our open-source firmware to build their own.
Web flasher
No drivers. No IDE. No esptool commands. Just plug a board into your laptop and click a button — PlantHub installs over WebSerial, the same standard your browser uses to talk to a 3D printer or Arduino.
Pick your hardware kit
Choose ESP32-C3, S3, or classic ESP32. Each preset auto-loads its wiring diagram.
See the wiring guide
Capacitive moisture, DHT22, BH1750 — labeled pin-by-pin so you can't go wrong.
Click Connect & flash
Browser uploads firmware in 30 seconds. No drivers, no Arduino IDE, no command line.
Who it's for
PlantHub works whether you've never touched a microcontroller or you're running a multi-zone greenhouse. Pick how you'd like us to introduce it.
PlantHub does the thinking. Plug in a board, claim it in the app, pick your plant — the AI handles watering, ventilation, and reminders. No YAML, no soldering, no plant-care guilt.
Use cases
Keep your balcony tomatoes and herbs thriving without daily check-ins. AI learns your garden’s rhythms and adjusts watering around the weather forecast.
Perfect humidity for your monstera. Automatic misting when needed. AI-tuned ranges for 50+ species out of the box.
Create named zones, assign devices, and cascade AI settings per zone. Override modes without touching individual plants. Zone-level location drives weather data.
Scale to dozens of devices. AWS IoT Core with mutual TLS in production. Flash a mock greenhouse profile to test the full pipeline before buying sensors.
Why you can trust it
A garden controller talks to water pumps and heaters. It has to be safer than a prototype. Here’s what holds PlantHub together when nobody’s watching.
Your data is scoped at the API, filter, database, and MQTT layers. Guessing another user’s device ID still gets you nothing.
Reasoning, sensor data observed, weather context, commands issued — all logged against the session that made them.
Max durations, per-actuator cooldowns, and daily evaluation caps live in the command pipeline — not in the prompt.
Mutual TLS to AWS IoT Core. Per-device X.509 certificates and scoped IoT policies. Each device can only talk to its own topics.
FAQ
Your rules keep running. PlantHub compiles the active schedule to on-device rules, and the broker’s last-will message lets the dashboard flag the device offline the instant it reconnects.
Yes. The internet is used for AI decisions, weather fetches, and app sync. Day-to-day watering and ventilation rules execute locally.
Yes. Each device has a chat interface where you can ask questions like "Why are my leaves yellowing?" — the AI uses your sensor history, weather data, and learned patterns to respond in plain English. Conversations are saved so you can pick up where you left off.
Adaptively. If your plant is stressed, the AI checks more often. If everything’s stable, it waits longer — saving energy and API costs. You can also set a fixed schedule or switch to fully autonomous mode.
Yes. Based on your sensor data and plant profile, the AI analyzes patterns and recommends threshold-based automation rules you can approve with one tap — no manual if-then configuration needed.
50+ profiles out of the box — tomatoes, peppers, herbs, succulents and more. You can create custom profiles with your own moisture, temp, and light ranges.
Yes. PlantHub firmware runs on ESP32-C3, S3, and classic ESP32 boards with standard I2C and analog inputs. Copy the built-in sensor template, wire your hardware, and the backend auto-detects whatever keys the firmware publishes — no server-side schema update needed.
Your tenant data is isolated at the API, filter, database, and MQTT layers — another user cannot query or subscribe to your devices. Every AI command is logged with full reasoning for your own audit.
Yes. PlantHub sends push notifications straight to your phone via AWS SNS. You choose exactly which events trigger alerts — threshold warnings, AI decisions, device-offline events, or all of the above.
Free during the beta. Post-launch pricing will be announced to waitlist members first, with early-bird rates locked in.
It’s on the hardware roadmap, no public ETA yet. USB-C is the supported power option today and will remain stable long-term.
If you have wifi and plants, PlantHub works. Weather data covers cities worldwide via OpenWeatherMap.
Yes. Flash one of three mock device profiles — Greenhouse Controller, Weather Station, or Hydroponic Garden — on a bare ESP32 board. They generate realistic sensor data every 30 seconds so you can test the full pipeline (provisioning, dashboard, AI, actuator commands) with zero wiring.
Six types out of the box: water pump, heater, shade blind, grow light, mist sprayer, and vent fan. Adding custom actuators follows the same template system as sensors — copy, paste, and include.
Early access
We're launching soon. Join the waitlist for early-bird pricing, first access to hardware, and input on features we build next.