st analyze

st-analyze — Analyze cross-product data and synthesize a new story

Generates a narrative summary of the cross-product fact-check results — who scored highest, where AIs disagreed, and which claims were consistently disputed.

Run after: st-cross

---

Examples

st-analyze subject.json                 # analyze and synthesize with default AI
st-analyze --agent gemini subject.json     # use a specific provider
st-analyze --no-cache subject.json      # bypass API cache
st-analyze --ai-caption subject.json    # AI caption of analysis results
st-analyze --ai-story subject.json      # full AI narrative from analysis

Options

Option	Description
file.json	Path to the JSON container
--agent {xai,…}	AI provider to use (default: gemini)
--cache	Enable API cache (default: enabled)
--no-cache	Disable API cache
--site {MMD,…}	Discourse site to use (default: MMD)
-v, --verbose	Verbose output
-q, --quiet	Minimal output

AI content generation

Option	Description
--ai-title	Generate a ≤10-word title → stdout
--ai-short	Generate a ≤80-word summary → stdout
--ai-caption	Generate a 100–160-word caption → stdout
--ai-summary	Generate a 120–200-word summary → stdout
--ai-story	Generate an 800–1200-word narrative → stdout

Related: Three Stages st-cross st-heatmap st-verdict

---

For developers

Flattens fact[] entries via mmd_data_analysis.get_flattened_fc_data() and passes the structured data to an AI for a prose summary.