把你的故事变成插画

# Lean Story-Illustration Pipeline v1
2026-06-12 由 spike（`../story-pipeline-prompts/`，14 文件 ~250KB）瘦身而来。spike 验证了配方但流程不适合批量：单故事 14 次 LLM 调用、最终产物 87% 是包装层、35%+ 跨场景逐字重复。本版把 LLM 轮次压到 **4 次**、文件压到 **6 个 ~25KB**，配方不变。
## 已验证的生图配方（不变）
```

点上方按钮生成本阶段产出

# 01 — Style Lock (multimodal, one-time per style)
Run once per visual style with a multimodal LLM. Input: 3–5 style reference images sharing one art direction. Output: one `style-lock.json` + a short review digest. Reusable across stories.
## Prompt
You are a senior visual development director. You will receive 3–5 style reference images. Extract a reusable **Style Lock**: the rules that make every future image feel like the same illustrated world.

点上方按钮生成本阶段产出

# 02 — Story Text → Sections → Scenes (text-first, all provenance tiers)
One pipeline stage, four sub-steps. **Every story is text-first**: scenes are picked from real paragraphs, never invented free-floating — this is what makes every image slot land on a `story_sections` anchor (`section_id`) with zero fuzzy matching at render time.
```
(A) story text          T1: 1 LLM call writes it | T2/T3: provided rights-cleared text

点上方按钮生成本阶段产出

# 03 — Production Boards (4 image-gen calls per story)
Four 16:9 boards form the production bible; their images are the **image references for every final scene generation**. Each template below is a fill-slot generation prompt: fill `{…}` from `style-lock.json` / `story.json`, send to the image model. Foundation boards (character, world) first, from style ref images; secondary boards (composition, lighting) after, attaching the character + world board images as additional references.
## Common rules (apply to all four boards)
**Produce the board image directly. Do NOT ask clarifying questions, do NOT request more detail, do NOT reply with text — generate the image from the specification given.** (gpt-image is chat-capable and will stall an unattended run by asking follow-ups otherwise — gap-bucket ② / story-v1/014.)

点上方按钮生成本阶段产出

# 05 — Compact Scene Prompts (one LLM call per story, all scenes)
Replaces spike's master-scene + per-scene final-builder + platform exporter. One call writes the compact prompt for **every** scene. Input: `story.json` + `style-lock.json` + `board-qa.json` `patch_instructions`. Output: `final-prompts.json` — each `prompt` string is what gets stored in `story_image_slots.prompt` and sent to the image model with the 4 boards attached.
## Prompt
You are a senior image-generation prompt writer. Using the story data, style lock, and QA patch instructions provided, write one **compact, self-contained** image-generation prompt per scene.

# 05 — Compact Scene Prompts (one LLM call per story, all scenes)
Replaces spike's master-scene + per-scene final-builder + platform exporter. One call writes the compact prompt for **every** scene. Input: `story.json` + `style-lock.json` + `board-qa.json` `patch_instructions`. Output: `final-prompts.json` — each `prompt` string is what gets stored in `story_image_slots.prompt` and sent to the image model with the 4 boards attached.
## Prompt
You are a senior image-generation prompt writer. Using the story data, style lock, and QA patch instructions provided, write one **compact, self-contained** image-generation prompt per scene.