TruEra's SaaS product supports local compute ingestion pathways for data; model ingestion is not currently supported by the SaaS product.
Local compute means computations are done locally with computed explanations sent to the TruEra server for aggregation and analysis. The primary advantage here resides in eliminating the need to replicate models and a user environment on the server, making implementation easier when executed where the model runs; for instance, on NB, Spark, etc. Local compute also supports third-party models. The disadvantage with a local setup is that you have to scale the compute; it does not scale automatically.