EpisodeLoom helps independent creators and small studios break literary screenplays into verified shot-by-shot visual prompts while maintaining long-horizon character continuity across serialized episodic stories.
A real sample showing how episodic scenes decompose into deterministic JSON ready for video generation engines.
// Story Bible constraints active: // Character: "Minh" (Detective, 34yo, worn brown trench coat, silver signet ring) // Location: "Old Saigon Alley, neon drizzle, 9:16 vertical ratio" [SCENE BEAT] Minh stops under the flickering red neon sign of the tea house. Rain trickles down his brimmed collar. He glances sharply over his shoulder as footsteps splash through puddles behind him. He slips his hand inside his coat, gripping the brass tape recorder.
{
"scene_id": "EP03_SC02",
"continuity_verified": true,
"shots": [
{
"shot_index": 1,
"duration_sec": 3.0,
"camera": "Medium close-up, low angle, slow push-in",
"lighting": "Moody cyberpunk rain, red neon rim light",
"diffusion_prompt": "Cinematic 9:16 vertical, 34yo Asian detective Minh in wet brown trench coat under buzzing neon sign, water droplets on wool, subtle head turn, 35mm film grain --ar 9:16",
"target_engine": "Kling_v2 / ComfyUI_Wan2.1"
}
]
}
Pairing deep reasoning with high-throughput structured extraction.
System builder and AI media practitioner exploring serialized storytelling workflows. Building EpisodeLoom out of firsthand frustration with character drift and manual prompt fragmentation when producing episodic vertical micro-dramas with AI diffusion tools.