LBM predicts saturation 11 min ahead. Reroutes signage, opens taps, moves staff — before the wait starts.
Outcome
2h saved per person. On 100k people at €100/day → €2.0M/day recovered, plus a festival that feels magical.
Same platform. Same open API. Same anonymous by design.
Invisible by design
You see nothing. You feel everything.
Planta sensors hide inside walls, ceilings, floors and street furniture. The audience keeps dancing. The workers keep working. The building does the thinking.
Rock in Rio Lisboa · 100,000 people / day
Zero queues. Zero incidents. Pure show.
Anonymous presence sensors read the crowd 24×/second. Bars, WCs, gates and staff move in sync with the flow.
LoRaWAN · in the wall
The size of a coin.
Edge · in the ceiling
Flush. Silent. Present.
You never see them
Flush-mounted, coin-sized nodes disappear into plaster, ceilings and street furniture.
You feel the result
The room breathes right. The queue vanishes. The kWh drops. The concert flows.
Anonymous by design
Presence and flow only. Identities destroyed at the hardware layer.
The connective tissue · one platform
How Planta connects everything.
Every sensor, every protocol, every context — merged into one Large Behavior Model your building can speak to.
Anonymous by designEdge-first
1 · Sense
Edge hubFusionFlow map
Large Behavior Model
Planta LBM
Fuses raw signals into flows — presence, comfort, energy, safety — and speaks them back to your building.
Live
Open APIActuator bus
Anonymous at the ADC layer · ephemeral IDs destroyed on device · nothing personal ever leaves the building.
3 · Serve
Homes
Ambient comfort, safety, energy — invisible.
Festivals
Move 100k people without queues.
Industry
Diagnose 80% of buildings that under-perform.
Heritage
Protect old walls without touching them.
Cities
Kinder streets, quieter blocks, faster flow.
Sense → anonymous signals from any protocol.Fuse → Large Behavior Model at the edge.Serve → your building, in plain language.
Sensor universe · all sensor classes · 144+ integrations
Every sensor family. One flow.
From €2 MEMS chips to LoRaWAN gateways, cameras and satellite links — every tile shows real hardware, the protocol Planta speaks, and how we connect all sensor types into one building flow.
144 of 144 sensors · anonymous by design · edge-first
click any tile
Real stacks · how these sensors combine
Rock in Rio Lisboa · 2024 & 2026
Crowd flow at 100 000 people/day
Problem — Bars, WCs and corridors saturate faster than staff can react. No cameras allowed.
Stack — Anonymous LoRaWAN presence + mmWave people counters at bottlenecks. Ubiquiti WiFi CSI as backup layer. Edge fusion writes a live pressure map; radio nudges routed to staff.
Outcome — Predictive alerts 8–15 min before saturation. Zero PII stored.
HORSE Aveiro (Renault Group)
Industrial building diagnosis
Problem — 80% of a facility under-performs silently — HVAC drift, filter loading, ghost loads.
Stack — Vibration + energy + air fusion. Modbus/BACnet bridge to legacy BMS, LoRaWAN indoor gateways for retrofit points. Anomalies routed to work-order draft.
Outcome — Root-cause diagnosis, not alarms. Filter change 3 weeks before failure.
Planta residential prototype · Portugal
Anonymous residential flow
Problem — Space should adapt to people without ever identifying them.
Stack — mmWave zones + edge depth + ambient air. Identities destroyed at the ADC layer. Everything orchestrated by the Large Behavior Model on a local hub.
Outcome — Zero cameras. Zero PII. Full presence-aware routines.
Remote sites · agriculture · construction
Off-grid critical monitoring
Problem — No mains, no LTE. Building still needs to phone home.
Stack — LilyGO T-Beam over LoRa mesh (Meshtastic) with Swarm satellite fallback for daily heartbeats. Solar-powered edge node reports structural + weather flow.
Outcome — Global uplink at <1 W. Structural and micro-climate telemetry.
Casa do Impacto · TRIGGERS 4.0
Wellbeing tuned to the person
Problem — Air, light and thermal comfort should adapt per person, not per building.
Stack — Per-room air + thermal + acoustic sensing, orchestrated with circadian lighting and TRV control. Opt-in personal flow layered on top.
Saving €2M a day is real. Getting your Sunday afternoon back is bigger. Physical AI is worth building because the world feels lighter to live in.
The supermarket with no line
You grab bread, milk, wine. You walk out. That was fifteen minutes of your Sunday, back in your pocket.
The festival that feels magical
You catch the headliner. You breathe. Signage moves before you do. The show is the whole night — not the queue.
The room that knows
Air freshens ten minutes before your meeting. Lights warm as the sun leaves. Nobody touched a switch.
The bank you barely notice
You sit. An advisor is already free. You leave sooner than you planned. You come back.
The city that breathes
Cleaner air on your street. Bins collected before they overflow. A neighbourhood that feels alive, not surveilled.
The queue tax
Every hour in line is money left behind.
Slide the values. Watch the recovered value. This is what happens when a building starts thinking in flows.
The math, in the open
1. Hours recovered = 100k × 2h = 200k hours
2. One attendee-hour is worth €100 ÷ 10h = €10.00
3. Value recovered per day = €2,000,000
4. Value recovered per event (1 day) = €2,000,000
Recovered value · per day
€2.00M
per event day, at 100k people & 2h saved per person.
Attendee-hours recovered200k h
Value per attendee€20.00
Over 1 event day€2.00M
Conservative model: value per hour = daily spend ÷ 10 waking hours. Real festivals recover concessions, bar throughput, safety costs, staff overtime and NPS on top of this — often 3–5× the number above.
Physical AI · Flow Solutions
80% of buildings don't work as they should.
We make buildings smart — with sensors, self-diagnosis and a building that talks back. One flow across LoRaWAN, Matter, Home Assistant and every BMS you already have.
Your building
Sense
Understand
From sensors to a building that talks back.
Respond
LoRaWAN gateway · Rock in Rio LisboaEdge sensorSensor fusion
01 · Sense
Any sensor. Any protocol.
Discreet LoRaWAN, Matter, Zigbee and BACnet devices across the building — new build or retrofit. We also read the HVAC, meters and BMS you already have.
02 · Understand
The building diagnoses itself.
Edge AI turns raw signals into flows. Comfort, energy, occupancy and structural behaviour read as one living system — not disconnected dashboards.
03 · Respond
You talk with the building.
Ask in natural language. The building answers, adjusts and predicts — grounded in AFI, our Theory of Everything for how physical space becomes intelligent.
