Some of AHI scan technologies are able to perform processing entirely within the user's device, without the need for processing on servers. This does come with the drawback that Machine Learning models must first need to be available on the user's device and can require a larger file size footprint for the app that otherwise would not be the case for more simple offerings.
AHI's solution for this is to allow such resources to be remotely hosted, instead of needing to be packaged with the App install bundle. This is particularly important for some App distribution stores, where there is a specific size limitation. It is also possible to embed these resources where remote asset download is not feasible, such as due to network limitations or other factors.
An advantage of remote resources solution is that updates to scan models can be rolled out without the user needing to reinstall the App (or App needing an AHI SDK version update), in most cases.