WHAT COREFEEDBACK IS
The front of the engine: inbox → snippets → clusters → insights you choose to promote into the layer. Separate what's loud from what's material before the roadmap locks.
HOW IT WORKS
Inbox & processing
Reviews land in Feedback Explorer; the NLP pipeline runs when you're ready. Readiness is visible — no pretending the data is complete.
Evidence (snippets)
Snippets are quotable, classifiable units — so claims are checkable. Longer reviews → more snippets → higher classification cost than one-liners; taxonomy-driven NLP (not generative slop). Credits on successful classification only; manual work doesn't spend credits.
Clusters — themes you can inspect
Candidate themes from snippets. Accepting a cluster commits it and uses credits on that path; browsing doesn't.
Insights — claims with custody
Statements tied to clusters for leadership or production. Accepting an insight stakes the narrative — metered on that path. Reject what doesn't hold.
CREDITS AT A GLANCE
Manual use is free. Credits: successful auto-classification, accepting generated clusters or insights.
| Action | Credits |
|---|---|
Automated snippet classification (taxonomy-driven; 3 of 9 taxonomies live — charged per snippet on success) | 1 cr |
Accept generated cluster | 2 cr |
Accept generated insight (per linked cluster) | 5 cr |
Numbers match CreditsCatalog in-app. Taxonomy set evolves — this page updates with releases.
ASK ON THIS SURFACE
Page-aware: tools match Feedback Explorer, Snippets, Clusters, Insights. Real snippet IDs in quotes. If NLP isn't ready, you get that — not a confident essay about your genre.
STOP STITCHING THE STORY FIVE MINUTES BEFORE THE MEETING
Join the waitlist for access. The CoreFeedback pilot program is closed.