About & Methodology

What WattNow actually uses

• • The appliance types visually detected in your photo.
• • Typical wattage ranges for comparable appliances available in Indonesia.
• • The operating status you confirm (active, off, or unknown).
• • The daily operating duration you estimate.
• • The electricity tariff you configure.

The formulas

daily kWh (min) = (min watts × quantity × hours) ÷ 1000
daily kWh (max) = (max watts × quantity × hours) ÷ 1000

daily cost   = daily kWh × tariff (Rp/kWh)
monthly cost = daily cost × 30
annual cost  = daily cost × 365

Because wattage is a range, every cost is shown as a range too.

Detection: two layers

A browser-side object detector (TensorFlow.js) runs locally on the video stream a few times per second to draw live bounding boxes. Generic pretrained models do not recognise air conditioners, sockets, routers, air purifiers or ceiling lights, so after you capture a frame a multimodal AI vision model analyses the still image to identify appliances the live detector missed. The detector layer is modular, so a custom-trained electronics model can replace it later.

Tariff

The default rate is an example, configurable estimate of Rp 1.700/kWh. PLN rates change over time and vary by customer category, so you can edit it on the results page at any time.

Limitations

Actual electricity consumption may differ depending on appliance model, age, operating mode, inverter technology, temperature settings and user behaviour. A camera cannot see whether an inverter AC is cycling at partial load, and it cannot read a nameplate that is not visible.

For exact figures, check your appliance energy labels, the model specification sheet, and your electricity meter readings.

Language we use

"Device detected" means the appliance appears in the photograph. "Likely active" means the system believes it may currently be operating. When operating state cannot be determined visually, WattNow shows "Unknown" and asks you to decide. Estimates are never presented as measurements.