the content pipeline
generate, review, publish.
the agent drafts content for your missions. you review it in the inbox. once approved, it's scheduled and published. rejected drafts go back to the agent with your feedback for iteration.
AI-generated drafts
the agent generates content grounded in your knowledge base and voice model. nothing goes out until it's reviewed — or until you trust the agent enough to publish directly.
idea fanout
one content idea becomes per-channel drafts automatically. the agent adapts format, length, and tone for twitter, linkedin, instagram, and more — from a single prompt.
inline editing & approval
review drafts in a rich editor. edit the text, then approve or reject. approved drafts move to scheduling. rejected drafts go back to the agent with your feedback.
voice consistency scoring
every draft gets a voice match score against your subject's trained model. flag off-brand content before it goes live.
scheduling & publishing
approved content is scheduled for optimal posting times or published immediately. the agent handles platform-specific formatting, media sizing, and API limits.
revision history
track every edit, approval, and rejection. see who changed what, when, and why. full audit trail for compliance-sensitive teams.
draft review
edit, approve, and schedule from one screen.
the split view shows the draft editor on the left and a live platform preview on the right. see exactly how your post will look on twitter before it goes live. voice match scoring flags off-brand content automatically.
faq
questions about content drafts
can the agent publish without my approval?
only if you set the mission's autonomy level to allow it. in supervised mode, every draft requires your approval. you control the dial.
how does idea fanout work?
call content_drafts_fanout_from_idea with a single idea. the agent generates separate drafts for each connected channel — adapting length, format, hashtags, and tone per platform.
what is the voice match score?
when a subject has a trained voice model, every generated draft is scored against it. a 94% match means the draft closely mirrors your established tone. low scores flag content that needs editing.