WThe full agentic pipeline: ingest → scrape → taxonomy → normalize → content → export. It fills what it can and routes exceptions to review.
Workflow
stages on6Toggle a stage off to run a partial pipeline: disabled stages are skipped and their exceptions never reach the review queue.
Pipeline
Each stage runs in order: toggle any off to skip it1Ingestion
Merge files · union by part number
100%
2Scrape
Pull fields from vendor product pages
100%
3Map to taxonomy
Assign Speedway part types
99%
4Normalize attributes
Map attributes to your schema
82%
5Content generation
65-char titles · descriptions · FAB
63%
6Export to loadsheet
Prepare the loadsheet
0%
Throughput
Parts moved through each active stageIngestion4,084
Scrape4,084
Map to taxonomy4,060
Normalize attributes3,942
Content generation2,610
Export to loadsheet0
Exceptions
Routed to manual reviewMap to taxonomyqueued24
Normalize attributesqueued118