LuxePlatforms
Rather just talk?(816) 477-6802hello@luxeplatforms.comKansas City · Building nationwide
IoT Platform · AgTech05 / 16

A cloud-native IoT platform to remotely monitor and control a grow tent, with AI plant-health vision

Run an entire grow room from your phone: sensors, lights, camera, and an AI agronomist.

A custom-branded platform that lets multiple growers remotely monitor and control one grow site (sensors, lights, fans, smart plugs, and a live camera) with hourly AI vision scans for plant health. A stateless on-site edge appliance handles the hardware while the brains stay in the cloud.

Deployed to AWS over public HTTPS

A look at the interface.

Representative UI — abstracted, never a client's real data.
🔒app.workspace.dev/overview
Studio
Overview
Analytics
Projects
Inbox3
Reports
Team
Settings
Storage68%
Workspace/Overview
Search…
Overview
Last 30 days · updated 2m ago
Last 30 daysNew report
Revenue
$128.4k
12.5%
Active users
8,642
4.1%
Conversion
3.84%
0.6%
Avg. session
4m 12s
8.2%
Performance
RevenueTarget
JanFebMarAprMayJun
Recent activity
Atlas migration
$24,800
Active
Onboarding flow
$11,250
Review
Billing revamp
$38,400
Active
Mobile parity
$9,600
Done
Dashboard / ops console · representative UI
Remote
Control your grow from anywhere
Live
1080p camera, any time
AI
Spots plant problems early
Role
Solo, backend, iOS, edge
Year
2026
Status
Deployed to AWS over public HTTPS
Category
IoT Platform · AgTech
Built with
  • FastAPI
  • TimescaleDB
  • MQTT
  • SwiftUI
  • Claude Vision
  • balenaOS
  • AWS
(01) The Challenge

The hardware controller has no usable cloud API and the camera lost remote P2P, so a no-on-site-device path was impossible. The solution was a minimal immutable edge appliance bridging Bluetooth and the camera's LAN, with a real automation goal: a rules engine where sensors, VPD, schedules, and AI insights drive actions with safety guards.

(02) The Approach

How it
works

The hardware has no cloud API and the camera lost remote P2P, so a minimal immutable edge appliance on the camera's LAN bridges video while everything else lives in the cloud. FastAPI + TimescaleDB ingests MQTT telemetry and exposes a command path and live WebSocket feed. An hourly cheap-model vision scan auto-escalates to a frontier model only when something looks off.

What I built

(03) Inside the build
01

Async telemetry + control backend

FastAPI with TimescaleDB hypertables ingests MQTT telemetry, auto-registers devices, and exposes a command path plus a WebSocket live feed, with role-based JWT auth.

02

Tiered AI plant-health vision

An hourly cheap-model scan auto-escalates to a frontier model at full resolution only when something looks off, budgeted at a few dollars per site per month.

03

Immutable edge appliance

A balenaOS box on the camera's LAN bridges video, and the backend reverse-proxies its stream behind short-lived tokens so growers get live video without any network setup.

(04) Highlights

The
receipts

  • Backend live on free-tier cloud over public HTTPS, all endpoints verified
  • Live 1080p camera relay through a token-gated proxy
  • Two-tier AI vision budgeted at ~$3–5/mo per site
  • Daily encrypted backups and 5-minute health alerting
← Previous
Marketplace for Teachers
Next →
Physics-Simulated Cooking Game