01
Safety Gate
Every topic is screened with Azure Content Safety before the agents run.
Azure AI Human Approval
Sign in with Keycloak to access the human-in-the-loop workflow: safety screening, writer draft generation, manual review, publisher polish, and relevance scoring.
Workflow Chain
01
Every topic is screened with Azure Content Safety before the agents run.
02
A writer agent creates the first detailed article draft for review.
03
Editors can rewrite, approve, or reject the draft before publication continues.
04
Approved content is refined into a cleaner final version for distribution.
05
Completed runs receive a post-publish relevance score so quality stays visible.