Icepack Documentation

Icepack provides a series of APIs allowing you to access high performance AI to solve your problems quickly and effectively.

If you’re looking for the documentation on how to use the APIs - you’ve come to the right place!

The production APIs currently support protobuf, a high performance serialisation layer for sending and receiving data where the payloads need to conform to certain standards. The standard that the payload needs to conform to is referred to as a schema. We use the words model and schema interchangeably when talking about the different APIs and the functions they perform.

While there are several advantages to using protobuf, one important feature is that it performs certain checks when attempting to serialise a potential payload. What does this mean for you? If you can serialise the message, then you probably have a valid model. The data inside the message could still be nonsense, but at least you’ll be getting another message back letting you know that Rubbish in = Rubbish out.

For those of you who would prefer to send models using JSON an OPEN-API V3 Specification is available on the Icepack Github account.