Configuration
| Field | Type | Required | Description |
|---|---|---|---|
| Mode | select | Yes | Routing mode: Classify, Keep, Drop, or Group |
| Rules | rule editor | When applicable | Conditions that determine which output path to follow |
| Filters | filter editor | When applicable | Filter criteria for keep/drop modes |
| Grouping | group editor | When applicable | Grouping configuration for group mode |
Modes
Classify
Sends each document to a specific output branch based on classification rules. Define conditions that match document properties or content, and each condition maps to a different output edge.Keep
Passes documents through only if they match the filter criteria. Documents that don’t match are dropped from the pipeline.Drop
The inverse of Keep. Drops documents that match the filter criteria and passes everything else through.Group
Groups related documents together before passing them downstream. Useful for batching documents by type, sender, or other properties.Inputs and outputs
Allowed inputs: Trigger nodes, route, parse, review. Output: Data passed to the matching output edge(s) based on the routing conditions.Credit cost
0 credits for rule-based routing. 0.5 credits per document when using text-matching classification.Related
Conditional routing guide
Set up conditional routing with examples
Parse action
Parse documents before routing on content
Extract action
Extract data after routing
Merge action
Merge routed branches back together