These NiFi processors allows you to process NiFi records through Vertex AI Gemini.
It uses batch inference and structured outputs to allow you:
- Perform a large number of inferences (one batch per flowfile, one inference per record)
- Force the returned records to conform to a schema that can be used downstream to process the results with record processors
It works in two steps, SubmitGeminiBatchRecord prepares and submits the batch request. PollGeminiBatchRecord then waits for that batch request to complete before outputting the results as a JSONL flowfile.
Examples of how these processors might be used can be found in the examples folder.