Skip to content
TH

AI vision for factories and yards

Counting, checking and warnings, done by cameras on your line.

We build AI vision systems and the operator software around them, installed on-site in your plant.

Talk to an engineerSee the solutions
  • Runs on-site
  • Keeps working offline
  • Operator screens in Thai

What cameras can take over

Three jobs, three answers

Warn

Is a forklift in that aisle right now?

Beacons light up the moment a forklift enters, so people on the walkway know to wait.

How it connects

From camera to action, all inside your plant

  1. 01

    Camera

    Your existing IP cameras, or new ones we install.

  2. 02

    AI on an on-site PC

    A model trained on footage from your own line.

  3. 03

    Operator screen

    Touch-screen software in Thai, with English alongside.

  4. 04

    Action

    Beacons, alarms, conveyor control, reports and exports.

  • Runs on a PC in your plant and keeps working without internet.
  • Uses your existing cameras and equipment where possible.

Solutions

Systems we build

Case study

Pipe counting on conveyors

Counts pipes and tubes as they leave the line, lot by lot.

How it works

Meter reading from cameras

Reads mechanical meters from the cameras already pointed at them.

How it works

Container latch check

Confirms all four latches are closed before a container leaves.

How it works

Kit completeness check

Checks every part is in the tray before a kit is packed.

How it works

Forklift aisle warning

Lights a warning beacon whenever a forklift is in a rack aisle.

How it works

Stockpile volume monitoring

Estimates how much sand, gravel and stone is left in each bay.

How it works

Case study

Pipe counting, in daily use at a manufacturing plant

In daily use at a manufacturing plant, Thailand

Pipes and tubes are counted by eye as they come off the conveyor. When the count drifts, lots go out short or over-filled, and nobody finds out until the customer does.

Read the case study
counted vs. true count
92 / 91
5½-minute reference recording, counted end to end by the production service.
per frame
9.6 ms
Detection and tracking, on the same recording.
on one PC
4 lines
4 camera streams at about 13 frames per second each, at 35% GPU load.
runs fully on-site
Offline
No internet connection is needed to count.

What you get

One team, from camera to operator screen

  • Operator software in Thai

    Touch-screen stations, dashboards, history and Excel or CSV export, with English alongside.

  • Installed on your site

    Runs on a PC in your plant, not in someone else's cloud.

  • Connected to your equipment

    Beacons, relays, PLCs, scanners and the cameras you already have.

  • Trained on your footage

    Every system is trained on footage from your own site before it goes live.

  • Support after go-live

    We stay on after launch to tune, retrain and fix.

How we work

How a project runs

  1. 01

    Site visit and footage

    We walk the line with you, agree what success looks like, and record footage.

  2. 02

    Train and try it on your footage

    We train the model and show you the results on your own video before anything is installed.

  3. 03

    Install and connect

    Cameras, PC and operator screens, linked to your beacons, relays or PLC.

  4. 04

    Go live and support

    We train your operators on site and keep tuning as conditions change.

Every system is trained on footage from your own site before it goes live.

About us

About Cloud Intex

Cloud Intex is based in Bangkok, Thailand. We build AI vision systems and the operator software around them for factories, warehouses and material yards, and install them on-site.

Have a count, check or warning you'd like a camera to handle?

Tell us about your line. An engineer will reply within one working day.

Talk to an engineer086-409-1663