Entity Enricher icon

Entity Enricher

Enrich entities with multi-model LLM fusion, multilingual output, and expertise-driven strategies

Enrichment → Batch Enrich Advanced

AI-generated

Summary

Batch enrich multiple entities in a single API call with advanced control over models, strategies, and output options using Entity Enricher's streaming API.

Inputs

  • Authentication — Select API Key or OAuth2 authentication method for connecting to Entity Enricher.
  • Connected To Name or ID — Choose or specify the organization and credential linked to the connection.
  • Schema Name or ID (required) — Select or specify the target schema that defines the enrichment output structure.
  • Model Names or IDs (required) — Select two or more LLM models for multi-model fusion or pick 'Auto — best model' for the organization's top-scoring single model.
  • Attachment IDs — Optional comma-separated attachment UUIDs to use as source material for enrichment, applied to every entity in the batch.
  • Upload Input Binary Files — Toggle to upload binary files from input items as attachments, which are used as source material for all entities in the batch.
  • Binary Fields to Upload — Comma-separated binary property names to upload; all binary files uploaded if left empty.
  • Delete Uploaded Attachments After Enrichment — Choose whether to delete attachments uploaded by this node after the enrichment finishes; does not delete attachments specified by Attachment IDs.
  • Language Names or IDs — Select one or more languages for multilingual fields. At least one language must be selected.
  • Strategy Name or ID — Select the enrichment strategy, with 'auto' letting the server choose based on the schema.
  • Classification Model Name or ID — Optional model for pre-flight entity type classification to prevent hallucination on mismatched entities.
  • Arbitration Model Name or ID — Model for LLM-based conflict resolution when merging multi-model results; leave empty for rule-based merging.
  • Web Search Name or ID — Enable or disable built-in web search for supported models; locked off if no selected model supports web search.
  • Advanced Options — Additional optional settings including response schema channel and strict structured output toggles.
  • Timeout (Ms) — Set the inactivity timeout in milliseconds; the job cancels if no progress event is received within this period.
  • Include Enrichment Metadata — Include metadata such as cost, tokens, fusion details, and record IDs alongside the enrichment result.

Output shape

a list of items, one per input entity, each containing the enriched entity data and optional metadata

Outputs for each input item an enrichment result extracted from streaming events with multi-model fusion support; includes error details for entities with no result. Attachments uploaded by the node are deleted if configured. The results are ordered by the input entity index.

Links

Discussion