Support

Work through these in order; the first two cause almost every failure.

1. Same Wi-Fi. The phone and the computer must be on the same network. Mobile data will not work, and many guest networks block devices from reaching each other ("client isolation").
2. Windows Firewall. The first time the desktop app runs, Windows asks whether to allow it on networks — choose Private networks. If that prompt was dismissed: Windows Security → Firewall & network protection → Allow an app through firewall → tick WattTrue under Private. Your Wi-Fi should also be set to a Private network (Settings → Network & internet → Wi-Fi → your network).
3. The right address. A computer can have several addresses and only one is reachable from the phone. The desktop lists them with its best guess first; if one fails, try the next. The app uses port 8787.
4. VPNs. A VPN on either device can hide the local network. Pause it while testing.

The desktop shows a 6-character code of letters and digits. It stops other devices on your network from adding readings to your run. The easiest way to enter it is to scan the QR code on the desktop with your phone's camera, which fills in the address and the code together.

If you type it: case and dashes don't matter, and the code never uses the look-alike characters 0, O, 1, I or L.

While WattTrue is measuring, Android shows a WattTrue notification. If it is missing, the measurement was stopped.

• Some manufacturers stop background apps aggressively. Allow WattTrue to run in the background — usually Settings → Apps → WattTrue → Battery → Unrestricted; the names vary by phone.
• Android 15 limits this kind of background work to about six hours a day. A charger test needs minutes, so this only matters if you measure all day.

Some phones report no current at all, and some only report it while the screen is off. Run the test with the screen locked. If the readings stay at zero, this phone does not publish its charging current, and WattTrue will not invent one.

Phones report current in microamps or milliamps, and do not say which. WattTrue infers the unit from the size of the readings and shows its assumption next to every figure. If it guessed wrong, switch the assumption — the raw number the phone reported is always shown unchanged, and the figures are recalculated from it.

WattTrue signs power by what the phone says is happening: positive means energy going into the battery, negative means it is being drained. Manufacturers use opposite signs in the raw reading, so the raw number can disagree — that is expected, and it is shown as reported.

Coupling losses are normal: WattTrue judges a pad against 45 % of its label rather than 70 %. If a pad is still far below, check alignment first — move the phone on the pad and test again — then remove a thick case. A pad that makes the phone pause charging repeatedly is usually running hot or badly aligned.

Phones slow their own charging above about 80 %, so a test up there measures the phone, not the charger. Let the battery run down below 80 % and test again; somewhere between 20 % and 70 % is ideal.

A verdict needs at least 20 readings — about 20 seconds — and the first few seconds after plugging in are left out, because that is the charger and phone negotiating, not charging. A minute or two gives a steadier answer.

The run came from WattTrue's preview or test tools rather than a phone's battery, or it carries no record of where it came from. WattTrue will not judge a charger on readings nobody measured.

iOS gives apps only the battery level, in 5 % steps, and whether the phone is charging. WattTrue times the steps and reports a charge rate, labelled as an estimate. It never shows watts for an iPhone, because none were measured.

iOS also pauses apps when the screen locks, so iPhone runs are taken with the screen on. Compare iPhone runs with each other, at the same brightness.

The desktop app saves each run as a folder containing meta.json and frames.jsonl, and shows where the folders are kept. Open the folder, or pick both files, in the run viewer. frames.jsonl alone also opens, but without meta.json the viewer cannot tell whether the readings were measured, so it gives no verdict.

The files are read by your browser and never uploaded.

The run was recorded by an older version whose format has since changed. The viewer shows how many lines failed and summarises only the ones that still read, rather than guessing at the rest.