Story workbench

Stories are the editorial layer between raw signals and generated content. A story is a curated collection of signals with an editorial angle — the “so what” that ties signals together into a narrative.

Why stories

Without stories, the pipeline generates content from whatever signals the scout found. Stories give you control:

• Curate which signals matter for a specific narrative
• Set an angle that frames the content direction
• Add editor notes per signal explaining why it matters
• Generate on demand from a specific story rather than the general signal pool

Creating a story

POST /api/stories
{
  "title": "Kubernetes security gets serious",
  "angle": "Three major CVEs in two weeks signal a turning point for K8s security practices",
  "editorial_notes": "Focus on practical takeaways, not fear. Our audience runs production clusters.",
  "signal_ids": [12, 15, 23]
}

Adding signals

Browse signals from the scout or SIGINT feed and attach them to a story:

POST /api/stories/{story_id}/signals
{
  "signal_id": 28,
  "editor_notes": "This Reddit thread shows real-world impact — use as evidence"
}

Each signal-story attachment has its own editor notes explaining its role in the narrative.

Discovering signals

AI-powered signal discovery finds additional relevant signals for a story:

POST /api/stories/{story_id}/discover
{"mode": "web"}

This searches the web based on the story’s title and angle, returning candidate signals you can attach.

Generating from a story

POST /api/stories/{story_id}/generate
{
  "channels": ["linkedin", "blog", "x_thread"],
  "team_member_id": 5
}

Generation uses:

• The story’s signals (not the general signal pool)
• The story’s angle as editorial direction
• Editor notes for each signal as context
• The org’s voice profile (or a specific team member’s voice)

Story workflow

Create story → Add signals → Set angle → Discover more → Generate → Review content

1. Spot a trend in the scout or feed
2. Create a story with a working title
3. Attach relevant signals with notes
4. Use discover to find more evidence
5. Refine the angle
6. Generate content for target channels
7. Review and approve in the content queue