Routing
Routing sends each processed document to the destination that matches it, with a fallback for anything else. Set it up with the Routing policy, or in any pipeline's output settings.
Set up the Routing policy#
- Open Processor → Pipelines and select Set up on Routing.
- Under Watch, choose the folder, bucket, or webhook the policy watches, or connect a new source.
- Add a route for each group of document types and choose its destination.
- Under Everything else goes to, choose the fallback destination.
- Select Create pipeline.
The Routing policy classifies each document first, so it needs AI classification; see Classification. To route without AI, select Customise and match on file properties instead.
Matching options#
| Match by | Values |
|---|---|
| File extension | pdf, docx, png, and other extensions |
| Exact filename | Complete filename |
| PDF title | Title metadata |
| PDF author | Author metadata |
| Document type (AI classification) | Labels produced by a Classify step |
Enter exact property values separated by commas. Wildcards and regular expressions are not supported. Only classification-based rules require AI.
Add rules in the pipeline builder#
- Create writable sources for your destinations.
- Enable routing in the pipeline's output settings.
- For each rule, select Match by, enter values, and choose a destination.
- Order the rules: the first match wins.
- Choose a fallback destination for unmatched files.
For classification-based routing, include a Classify step before delivery.
Example without AI#
| Order | Condition | Destination |
|---|---|---|
| 1 | File extension is pdf |
PDF archive |
| 2 | File extension is png, jpg, jpeg |
Image archive |
| Fallback | No rule matches | Unsorted |
Test both matching and unmatched files. If an earlier step converts the document, check that your rules match the output format.
