How it works. It connects your Octopus Energy account, pulls your half-hourly consumption data, finds your location, and downloads the matching solar generation profile for your postcode. The rest is maths.
Limitations. It uses historical data to model the past year as if you had solar/battery, is also currently uses 2024 pricing. I think this is a reasonable approach as nobody knows what the future holds, My only goal is to say "based on what happened in the last year, this is how much you could have saved".
About the battery model. The battery is fully optimised, and dispatched with perfect foresight to maximise revenue for the homeowner. Low value trades reduced by varying the avoid skip trade threshold slider. This means that the battery will not dispatch unless the gain from dispatching is above a certain price. When this is not set the battery finds the maximum a real battery could earn at the chosen tariff. Octopus free-energy windows and "saving session" incentives are not included. A few battery parameters (max charge/discharge rate, wear threshold) can be adjusted from the Battery settings panel once results load; efficiency and SoC limits are fixed at sensible defaults.
Heat pump scenarios. If your Octopus account includes gas data, the app can estimate your design heat loss from winter gas use and local weather. You can also choose a heat-loss value manually; the app then matches a measured HeatPumpMonitor system and adds its electricity demand before re-running the solar and battery results.
Your data. The API key is used in-memory once per session. It is not stored, logged, or shared. You can also try the app without an Octopus account using the "Test without my data" button.
What you can change. Solar array size in kWp, tariff (Agile, Cosy, Go, or flat), and the battery parameters mentioned above.