monitorMyLawn
Sign up

How it works

Ten minutes to install. About a week before it has much to say.

Buy it, push it in, point it at the internet, claim it — that part really is a ten-minute job with no trenching, no wiring and no hub. Then comes the part most of this industry skips over: the tools are fits over your own history, and for a few days there is not much history to fit. Here is exactly what works from the first uplink, and what genuinely does not.

Start monitoring What the eight tools do

No proprietary hubNothing wired to your valves Wi-Fi, MQTT or LoRaWANChangeable per device

Section through turf and soil showing where the probe has to reach. A cutaway from the grass canopy down to eight inches of soil, with a ruler marked every two inches. The top inch is thatch, shaded pale because it dries within hours of a sprinkler finishing. Below it the root zone, three to six inches down, is shaded dark because it holds water for days. One probe has been pushed in only far enough to leave its sensing section in the thatch and is marked wrong; the other reaches the root zone and is marked correct. 0 IN 2 IN 4 IN 6 IN 8 IN THATCH · DRIES IN HOURS ROOT ZONE · HOLDS FOR DAYS IN THE THATCH IN THE ROOT ZONE
Push until the sensing section is past the thatch. A probe sitting an inch too high reports a thirsty lawn every afternoon — confidently, and for months.
FIG.01

Buy

A stake, and a subscription for the stake.

Two separate things, priced separately, and we would rather say so plainly than bury it. The hardware is a one-time purchase. The subscription is what keeps that device allowed to send readings and keeps its history alive.

What is in the box, and what is not

There is deliberately no hub, no bridge and no base station. The stake talks to the internet on its own, which is why the connection choice below is yours rather than ours. Our reference ESP32 firmware ships with the hardware — it is not a download from this site, so do not go looking for one.

  • Channels8
  • Soil moisture%VWC, 0–100
  • Uplink optionsWi-Fi HTTPS · MQTT · LoRaWAN
  • Hub requirednone, ever
  • Sample intervalset on the device · 5–15 min typical
  • Ingest ceiling30 readings / min / device
  • Subscription$4.99 / month or $49 / year, per device
  • Battery & lifeto confirm
  • Ingress ratingIP__ · to confirm
  • Wireless rangeto confirm
  • Probe depthto confirm

The rows marked to confirm are hardware facts nobody has verified on a shipping unit yet. We would rather show a gap than a number we would have to retract.

How many do I need?

One per area that is watered differently or behaves differently. A single front lawn on one sprinkler zone is one sensor. A front and back on separate zones want one each. A shady strip that stays damp for days wants its own — because averaging it with the sunny corner hides the exact problem you are trying to find.

Billing, in one paragraph

One account, one subscription, whose quantity simply tracks how many sensors are active. Add a sensor and it goes up; deactivate one and it comes back down on the next invoice. Cancel any month from the billing portal — no phone call, no retention script. A sensor that is not paid for stops sending new readings, but everything it already sent stays visible and stays exportable.

Fig. 01 — What arrives and what it costs. Hardware once, subscription monthly or yearly, and every agronomy tool included at both tiers.

FIG.02

Stake

Where you put it decides what it can tell you.

This is the one step you can get wrong in a way no amount of maths will fix. A probe in the wrong ten square feet reports a lawn you do not have — and it will report it confidently, for months.

Plan view · one property · three zones Good Avoid
Plan view of a property showing good and poor sensor placements. A house sits at the top with planting beds along its front wall and a downspout at one corner. A driveway runs down the right-hand side. Three lawn areas are marked: a sunny front lawn, a shaded side strip, and a back lawn. Three good placements are marked in the open middle of each lawn area, away from edges. Four placements to avoid are marked: hard against a bed edge, directly under the downspout outfall, in the low corner that puddles after rain, and in the strip beside the driveway that gets reflected heat. HOUSE PLANTING BEDS DRIVEWAY SHADED SIDE STRIP DOWNSPOUT PUDDLES AFTER RAIN FRONT LAWN · FULL SUN 1 2 3 A B C D
Good — open middle of each zone
  • 1 — Front lawn, full sun. Out in the open, well away from the bed edge and the drive. This is the grass you look at, so this is the grass to measure.
  • 2 — The shaded side strip. Its own sensor, because it stays damp for days and averaging it with the front lawn produces a number describing neither.
  • 3 — Back of the sunny area. Far enough from the driveway to escape the reflected heat, and on the same sprinkler zone as the head that covers it.
Avoid — and why
  • A — Hard against a bed edge. Bed irrigation, bed mulch and root competition all bleed in. You would be measuring the border, not the lawn.
  • B — Under the downspout. Every shower reads as a flood, every dry spell as a drought, and the leak detector will have opinions.
  • C — The corner that puddles. Standing water pins the probe near saturation and the advisor will happily tell you never to water again.
  • D — The strip beside the driveway. Reflected heat and a hard edge dry it faster than anywhere else, so it reports a thirst the rest of the lawn does not have.
Depth matters more than position

Push until the sensing section sits in the root zone, not in the top inch of thatch. Thatch dries out within hours of a sprinkler finishing while the roots below it are still comfortable for days — a probe sitting too high will report a thirsty lawn every single afternoon and have you watering a lawn that does not need it.

Fig. 02 — A representative property, not yours. The rule that transfers is the one about edges: every boundary in a garden — bed, path, drive, downspout — is a place where the soil stops behaving like the lawn.

FIG.03

Connect

Three routes. Pick one per device, change your mind later.

This is where most soil sensors quietly insist on their own hub, in range of their own radio, powered forever. We never require one. All three routes below end at the same validated, de-duplicated ingest pipeline and the same dashboard.

Wi-Fi, over signed HTTPS

The lawn is inside your house Wi-Fi. This is the simplest possible setup and needs nothing else running.

  • What it needsRouter coverage at the exact spot the stake will live. Walk out there with a phone and check before you commit — a bar of signal at the back door is not a bar of signal at the fence.
  • How it authenticatesEvery request carries an HMAC signature and a timestamp. A reading cannot be forged, and it cannot be replayed more than five minutes later.
  • What setup hands youThe ingest endpoint, your key, and a copy-paste request-signing snippet.

MQTT, to a broker you control

You already run a broker or a home-automation stack and would rather have the sensor on the same fabric as everything else.

  • What it needsA broker you control. We issue per-device credentials that are locked to that device’s own topic, so one sensor can neither read nor write another’s.
  • How it authenticatesPer-device broker credentials, scoped to one topic. The device publishes the same metrics object as the HTTPS route.
  • What setup hands youYour topic, your credentials, and a publish example.

LoRaWAN, through The Things Network

A paddock, an allotment, a far corner, a rental. It trades bandwidth for range, and a lawn sensor needs almost no bandwidth.

  • What it needsA Things Network gateway in range — often one that already exists and is not yours. This is the answer when Wi-Fi simply does not reach.
  • How it authenticatesNetwork-session keys on the Things Network side, plus a bearer token on the webhook that forwards uplinks to us.
  • What setup hands youThe complete Things Network payload formatter, ready to paste into the console.
The same eight numbers, whichever route POST /api/ingest.php 
POST /api/ingest.php
X-MML-Key:       mmlk_…
X-MML-Signature: t=<unix-ts>,v1=<hmac-sha256-hex>

{
  "device_uid":  "mml_8f3ad21c",
  "recorded_at": "2026-06-21T14:30:05Z",
  "metrics": {
    "soil_moisture":    42.7,   // %VWC
    "soil_temperature": 19.4,   // °C, canonical
    "air_temperature":  24.1,
    "air_humidity":     55.0,
    "light":            38400,  // lux
    "rain":             0.0,    // mm
    "battery_voltage":  4.02,
    "signal_strength":  -67     // dBm
  }
}

MQTT devices publish that same metrics object to mml/<device_uid>/readings. LoRaWAN nodes send nothing like it — The Things Network posts its own uplink shape to a separate endpoint and our payload formatter maps the decoded fields onto the same eight keys, with the gateway’s RSSI becoming signal_strength. You never write that mapping yourself; it is generated for you on the claim screen.

  • How often should it report? Your choice, set on the device. Every 5 to 15 minutes is typical and comfortably inside our limits. Reporting more often costs battery and tells you very little extra, because soil moisture moves slowly.
  • If it drops out Readings buffered on the device are accepted when it reconnects, and a late reading will not rewind your last-seen time. Nothing for three hours raises an offline alert and emails you.
  • Secrets Device secrets are stored only as hashes and shown to you exactly once, at the moment they are issued. If you lose one, you rotate it rather than recover it.

Fig. 03 — A mixed fleet on one account is fine: Wi-Fi at the house, LoRaWAN at the far paddock, all on the same dashboard. Nothing in this section is a product we sell you twice.

FIG.04

Claim

Ten minutes, most of it copy and paste.

Claiming is what ties a piece of hardware to your account and issues its credentials. It is the only fiddly part, and the setup screen is built to hand you every string you need rather than describe it.

  1. Step one · about a minute

    Create an account and enter the device ID and claim code printed on the hardware. Both are on the unit itself, so a device you are holding is a device you can claim.

  2. Step two · a dropdown

    Choose the connection type — Wi-Fi, MQTT or LoRaWAN. This is per device, and it can be changed later without re-claiming anything.

  3. Step three · shown once

    We issue a signing key and secret. They are displayed exactly once and stored only as hashes, so nobody — including us — can read them back to you afterwards. Copy them now; if you lose them, rotate rather than recover.

  4. Step four · the agronomy inputs

    Set soil type, grass type, zone area and sprinkler precipitation rate. These are what the advisor computes against — leave them at their defaults and it assumes loam, cool-season grass and a slow rotor at 0.5 in/hr. It is worth five minutes and a catch-can test to get the rate right.

  5. Step five · the bit that feels good

    Paste the credentials onto the device. The dashboard watches for the first reading and tells you the moment it lands, so you are never left refreshing and wondering whether it worked.

What the claim screen actually gives you

Your ingest endpoint, your key, the exact request-signing snippet, an MQTT publish example, and the complete Things Network payload formatter for LoRaWAN nodes. Our reference ESP32 firmware ships with the hardware rather than as a download from this site.

Fig. 04 — The device moves from unclaimed to pending to active. Ingest is only accepted once the device is active, which is also the moment it starts counting toward your subscription quantity.

FIG.05

Watch

The first week, honestly.

Live readings and threshold alerts work from the first uplink. Most of the interesting tools do not, because they are fits over your history and on day one there is no history to fit. A product that pretended otherwise would be giving you a confident number computed from three data points. Below is when each one actually starts earning its place.

When each tool starts saying something worth acting on
  • Answers, but thin
  • Trustworthy
A timeline of when each tool becomes useful after a sensor is installed. Live readings, built-in threshold alerts and mow, fertilize and seed readiness are all trustworthy from the first uplink. The Lawn Health Score computes from the first uplink but is only fully trustworthy after a week. Anomaly and leak detection begins after about an hour, once three readings exist, and is complete by day one when twelve samples are available for flatline detection. The sensor-offline alert can first fire at three hours. The Smart Watering Advisor becomes reliable at day one, once a full day and night temperature swing has been recorded. The Dry-down Forecast needs until day two, when a moisture peak and a falling segment exist. Infiltration Efficiency starts at your first logged cycle around day two and needs four cycles, so it is not trustworthy until week two. Water Savings needs about a week of logged cycles and is only meaningful from week two. Multi-device Compare is not time-gated at all — it needs a second sensor. First uplink +1 h +3 h Day 1 Day 2 Day 3 Week 1 Week 2+ Live readings, all eight channels Charted the moment the first uplink lands. Built-in threshold alerts Low moisture, battery, frost and heat are checked on every single reading. Mow, fertilize & seed readiness Needs soil temperature, which arrives with the first reading. Lawn Health Score Computes at once, but stability needs a spread and light needs whole days. Anomaly & leak detection Three readings to run at all; 12 for flatline, 8 for spikes, 6 for the battery fit. Sensor-offline alert Three hours of silence, by definition. It cannot be faster than that. Smart Watering Advisor Needs one day–night swing so Tmax and Tmin are real numbers. Dry-down Forecast Needs a moisture peak and at least three samples falling away from it. Infiltration Efficiency Starts at your first logged cycle; the trend needs four of them. Water Savings Needs a window of logged cycles long enough to be worth comparing. Multi-device Compare Not time-gated. It needs a second sensor, not more days. ADD A SECOND SENSOR
Nothing on this chart is throttled by us. Each bar starts where the maths genuinely becomes valid — three readings for the anomaly scan, twelve for flatline detection, one day–night swing for ET₀, a moisture peak plus three falling samples for the forecast. Below its threshold each tool returns a plain-English diagnostic rather than a number.
  • Day one, realistically You will watch the numbers far too often, and the most interesting thing will be how little soil moisture moves. That is the correct lesson: soil is slow, which is exactly why a daily timer is the wrong instrument.
  • The first dry-down is the milestone Once you have watched moisture fall from a peak for two or three days, the forecast has a slope of your own soil to work from — and from then on it is describing your lawn rather than a generic one.
  • Log your cycles Infiltration efficiency and water savings both need to know when you watered and how much. Log a cycle from the dashboard, or the advisor’s own recommendation records it for you. Unlogged hand-watering can also read as a leak.

Fig. 05 — The honest onboarding curve. Every threshold on it is a real guard in the code rather than a marketing decision, and each is listed per tool on the tools page.

FIG.06

Act

It tells you. You turn the tap.

monitorMyLawn does not control your sprinklers, and that is a deliberate choice rather than a missing feature. It keeps the product compatible with every irrigation setup there is — including a hose and a mechanical timer — instead of only working with one brand of smart valve.

What arrives, and when
  • A sentence, not a spreadsheet. Water or hold, a run time in minutes, and a pre-dawn start — with the ET₀, the rain and the depletion that produced it printed underneath so you can disagree with it.
  • Alerts when something is wrong, in the dashboard and by email, with quiet hours so nothing wakes you at three in the morning.
  • An optional weekly digest on whichever day you choose, instead of a stream of individual messages.
  • Silence the rest of the time. An alert that fires constantly is one you learn to ignore, so the thresholds clear at a different number than they raise at.
What you can do with it afterwards
  • Log the cycle you ran so infiltration efficiency and water savings have something real to measure. The advisor’s own recommendations log themselves.
  • Send a read-only link to a landscaper, groundskeeper or HOA. It carries no personal information and you can revoke it instantly.
  • Export everything as CSV per sensor, in your display units or the raw canonical ones, up to two years and 400,000 rows a download — plus a full account export in JSON.
  • Change your mind. Cancel any month from the billing portal, or deactivate one sensor and keep the rest running. Your history stays readable and exportable either way.

Fig. 06 — The only thing that changes about your irrigation is when you turn it on, and for how long.

Put one in the ground

Stake it in the corner that browns first.

Inside a day you will know how fast your soil actually dries. By the weekend you will know how many minutes it needs. That is the whole product, and the install really is ten minutes.

Start monitoring — $4.99/device/mo Ask us about your yard first

Not sure how many you need? Send us the layout and we will tell you honestly — including if the answer is one.

  • No hubNothing else to buy, power or wire into your sprinkler valves. The stake reaches the internet on its own.
  • No lock-inConnection type is per device and changeable, and every reading exports as CSV or JSON.
  • No pretendingEach tool says how much history it needs, and returns a diagnostic instead of a number below it.
  • No surprises$4.99 per device per month. Cancel any month, keep your history.