AI Pulse Analysis

Pulse analyzes each scene with its previous and next scenes, then saves grades and editorial feedback in scene properties for hover review.

• Manuscript or subplot order: Evaluate the reading sequence or follow one subplot.
• Scene feedback: Review strengths, continuity, and opportunities to improve scene flow.
• Compact hover: Turn off Pulse context in Settings → AI to show the current scene’s feedback alone.

Modes: Progress mode (key 1), Narrative mode (key 2), Chronologue mode (key 3) Command: Scene pulse analysis (manuscript order), Scene pulse analysis (subplot order) Settings: Settings → AI — enable Enable AI LLM features and choose a provider

_{AI Pulse Triplet Analysis}

_{Pulse triplet output — previous, current, and next scene grades surfaced in hover}

Supported Providers

Pulse works with all supported AI providers:

• Anthropic Claude
• OpenAI GPT
• Google Gemini
• Local LLM (Ollama, LM Studio, and OpenAI-compatible servers)

Local LLM setup is documented under Settings → AI → Local LLM. Local runs are validated before analysis starts and results are written with the same safeguards as hosted providers.

When a Pulse batch finishes, its card shows the batch’s total cost. Cache use and per-scene costs are under AI prompt & context.

For the command-specific batch workflows, see:

• Scene pulse analysis (manuscript order)
• Scene pulse analysis (subplot order)
• Summary scene refresh