Scene Plan Prompt

You are creating a scene plan for a vertical short-form video.

Input:
- Script JSON: {{script_json}}
- Visual style: {{visual_style}}
- Safety: {{safety}}

Requirements:
1. Create one visual scene per narration segment.
2. Every scene must have a distinct visual focus, camera angle, and motion idea.
3. Prompts must forbid visible text: no letters, numbers, subtitles, captions, logos, labels, signs, watermarks, fake writing, or decorative glyphs.
4. Preserve continuity across scenes: era, palette, location logic, and character descriptions.
5. Do not add dialogue or generated speech to image-to-video prompts.

Return JSON with:
- scenes: [{"id", "segment_id", "visual_focus", "camera", "image_prompt", "motion_prompt", "duration_seconds"}]
- continuity_notes
- text_free_guard