Ask HN: Would you pay for a forecasting API?

It is simple as follows:

You upload your time-series data (e.g. csv) that contains data points (contains arbitrary number of features and y column as the main value) from a timespan T1-T2 (monthly, daily, sub-daily, etc.). We return you the predicted values for the timespan T2-T3

Forecasting/prediction might be trained daily or near-realtime fashion.

So far we had worked on fast-food transactions over the time and found accurate results in daily and hourly basis.

Should we work on this further and provide it as a service? Would you pay for it?