Skip to content

cassIO, general architecture

The purpose of CassIO is to abstract away the details of accessing the Cassandra database for the typical needs of Generative AI or other Machine Learning workloads, so as to offer a low-boilerplate, ready-to-use set of tools for seamless integration of Cassandra in most AI-oriented applications.

CassIO is framework-agnostic: it does not know the specific implementation details of interfaces of LangChain, Llamaindex or any of the various toolkits built around Generative AI. Rather, for each such framework, a set of "thin adapters" is built that fits the interfaces of the framework while leveraging the capabilities implemented in CassIO.

Here is a sketch of the relation between your application, an LLM framework such as LangChain, and CassIO:

CassIO, sketch

Latest status

CassIO is evolving rapidly. Note that at the time of writing

  1. only LangChain is supported;
  2. a fork of LangChain is needed (a PR to upstream is on its way).

As long as you follow the setup instructions we offer, this is easy and quick.