Does Ingest Node provide any value if not using pipelines?

We are using our own ingestion process of highly structured data (many tables from multiple databases). I wonder if ingest node in general provide any low level capabilities to be useful for custom ingestion process.

I would greatly appreciate your opinion

At this stage it is just a subset of basic Logstash functions - grok, geoip etc. So it is unlikely to be useful for you.

Perhaps the simulate api can be useful to you? You can set it up to run with grok, geoip or attachment processor and the output you can then use in your ingestion process?