Know the moment today stops looking like this plant's normal day.

The problem

A single out-of-range reading — a slightly high VPD, a slower growth day — rarely means much on its own, but a cluster of small shifts across several signals on the same day usually does. Watching a dozen separate numbers on a dozen separate screens, and noticing when several of them drift at once, is exactly the kind of pattern a person tracking by hand tends to miss.

Why it's hard to solve alone

The earliest sign of a developing problem is often not one dramatic number but several ordinary ones moving together in the same direction — humidity edging up, growth slowing slightly, leaf color drifting, all in the same few days. A comparison against this specific plant's own recent history catches that kind of drift far earlier than any single fixed threshold could, because 'normal' here means normal for this plant, this stage, this cycle — not a generic target pulled from somewhere else.

How GrowScope helps

GrowScope builds a rolling baseline from this plant's own recent days at its current growth stage — a stage change is treated as a new normal, not an anomaly — and scores today against that baseline using robust statistics that don't let one bad sensor reading or one skipped watering distort the whole picture. Any signal without enough reliable data simply doesn't count toward the score, and until there's enough history to trust, GrowScope says so plainly instead of reporting a false 'all normal.'

How it works

• GrowScope pulls together up to nineteen signals it already tracks — temperature, humidity, VPD, PPFD, DLI, CO₂, watering interval and volume, runoff, EC, pH, growth velocity, canopy temperature, leaf greenness, airflow uniformity and leaf motion, fan vibration — wherever you've actually recorded them.
• A rolling baseline is built only from this plant's own recent days at its current growth stage, so moving from veg to flower is treated as a new normal, not a deviation from the old one.
• Deviation is scored with robust statistics — the median and typical spread, not a plain average — so one bad sensor reading or one skipped log entry doesn't quietly poison the whole baseline.
• Until at least five days of the current stage are on record, the fingerprint honestly reports that it's still learning rather than reporting a false 'all normal.'
• The whole calculation runs on your device with no AI credits spent and no external model involved — the same deterministic-math principle GrowScope uses for its VPD and CO₂ targets.

What you get

You get one daily read on whether today looks like this plant's normal — built from everything else GrowScope already records for it, scored against its own history, at no cost and with nothing sent anywhere.

Frequently asked questions

Related features