Plug & Play AI Camera - 1
Plug & Play AI Camera - 2

Plug & Play AI Camera

One AI camera, a gallery of apps — install object detection, text reading, face analysis or people counting from the camera's own console, and forward the results to Home Assistant

初級20minオートメーション
recamerarecamera-prohome-assistantmqttapp-galleryyoloocrface-analysisfall-detectionretail-analyticssmart-homertsponvifprivacy

What This Solution Does

Plug in an AI camera, open its console in a browser, and install an app: object detection, text reading, face analysis, drowsiness monitoring, weather classification, QR scanning, rep counting, fall detection or retail people counting. Switching apps is a click. All AI runs on the camera, so video never leaves your network — and when you want the results elsewhere, the camera publishes them over MQTT and Home Assistant picks them up automatically.

Key Benefits

BenefitDetails
An app gallery on the cameraInstall and switch apps from the browser. Your computer does the downloading, so the camera itself needs no internet
One camera, many jobsThe same hardware detects objects, reads text, analyzes faces or counts customers, depending on which app is running
Home Assistant without glue codeThe camera publishes MQTT discovery configs; entities appear on their own, and change when you switch apps
Privacy by defaultAI runs on the camera. Video and detections stay on your local network, and the console can blur faces before the frame is even encoded
Works with what you already haveRTSP for any player, ONVIF discovery for an NVR or VMS (reCamera)

The Two Presets

PresetHardwareWhere apps come fromMQTT broker
reCamerareCamera 2002The reCamera Console's app gallery, installed by this solutionOn the camera, plus the one you deploy with Home Assistant
reCamera ProreCamera ProThe camera's own App Center, delivered by firmwareNone on the camera — the broker comes from the Home Assistant step

Both follow the same five steps: get the device current, pick an app, deploy Home Assistant if you want it, point both ends at the same broker, then build the dashboard.

Apps in the Gallery

AppWhat it does
Object DetectionPeople, vehicles and 80 other everyday classes
OCR Text ReaderChinese and English text from signs, labels and meter displays
Face AnalysisAge, gender and emotion, with optional privacy blur on the stream
Drowsiness DetectionEye closure, PERCLOS and yawn frequency from FaceMesh
Weather ClassificationClear / cloudy / foggy / rainy / snowy from the camera view
QR Code ReaderEvery code in frame decoded at once, no model and no handheld scanner
Fitness TrainerRep counting and form flags for squats, push-ups and curls
Fall DetectionRapid falls from an on-device pose timeline, for fixed indoor views
Retail People CountingEntry / exit counts plus browsing, engaged and assistance states

The reCamera Pro's App Center carries its own build of most of these, plus voice transcription. Weather classification is reCamera only.

Requirements

Network

  • The camera and, if you use it, the Home Assistant machine must be on the same local network
  • USB works for initial setup (address 192.168.42.1)
  • The computer running this app needs internet access to download apps on the camera's behalf; the camera does not

For the Home Assistant steps

  • Docker on the machine that will run Home Assistant, with ports 8123 and 1883 free
  • Or an existing Home Assistant and MQTT broker, in which case steps 3 and 4 collapse into filling in one address

Privacy

  • All AI detection runs locally on the camera
  • Video streams and detection data stay on your local network

デプロイ構成

camera
recamera_2002
server
recomputer_r11
お問い合わせ
ハードウェアパートナーとしてうれしいです!
次へ