Novel Storyboard logo

Novel Storyboard

CommunityPopular
eternityspring
novel-storyboard

给 AI 短剧出分镜:三层结构——段(一次视频生成,≤15 秒)→ 分镜(段内 2–5 秒的剪切,认领剧本节拍) → 分镜图(每切一张关键帧:主分镜图钉 0.00 秒,子分镜图钉各自切点)。 每段自带一条 MiniMax H3 视频提示词(官方口径默认英文、逐镜换行,promptLang 可切中文):对齐指令和 [Shot k] 切点时刻由分镜结构推导、逐字对账,台词逐字进 <d> 块(写法规范已内化为 references/h3-prompt.md,不依赖外部 skill)。 产出 storyboard.json + Markdown + 单页评审报告(分镜节奏带 / 分集分镜表 / 生成批次单 / 配音对齐单,含导出 JSON)。分镜图出图拿场景与角色设定图当参考图走 codex $imagegen(可选)。 17 道质量门全部由脚本确定性检查(第 17 道 shot-recipe 可选:挂上 shot-recipes 卡库才查,不挂就明说跳过); export 一键导出 H3 投产包(每段提示词 + 按 Picture 序的分镜图清单)。零依赖、零 API key,用当前会话额度。 Use when asked to 分镜、出分镜、镜头表、切镜、storyboard for AI short drama。

Overview

Publishereternityspring
Repositoryshuohao-skills
Skill namenovel-storyboard
Stars
3.5K
Forks
476
Bundled files
13
LicenseApache-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 13 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by eternityspring on GitHub. Read the source before you install it.

Installation

Install the Novel Storyboard AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/eternityspring/shuohao-skills.git /tmp/shuohao-skills
mkdir -p .claude/skills
cp -r /tmp/shuohao-skills/skills/novel-storyboard .claude/skills/novel-storyboard
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Novel Storyboard in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Novel Storyboard on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Novel Storyboard is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

novel-storyboard

给 AI 短剧出分镜——管线里第一个直接面对视频模型的层。前提刻在骨子里:镜头是生成出来的,多切一镜的成本几乎为零,所以这里不心疼镜头数量,上限只有一个:视频模型单段生成的时长(默认 15 秒)。

核心机制:镜头认领节拍。 每个镜头声明它覆盖剧本某场的哪几个连续节拍(sceneIndex + beats: [起, 止]),镜头不许跨场次——换景必换镜。这让分镜和剧本的关系变成可机械对账的:

交付解决什么
节拍认领每个节拍被恰好一个镜头认领、顺序不乱——剧本改了重跑 validate,失效的镜头当场点名
单镜头 ≤ 15 秒AI 视频单段生成上限,长对话在这里被强制拆镜(params.maxShotSeconds 按模型改)
台词装得下认领节拍的台词秒数 ≤ 镜头秒数——逐镜检查,不是拍脑袋
首帧 + 运动双提示词首帧给图像模型(配合参考图),运动是模型无关的过程描述;景别、运镜是枚举,英文短语必须写进对应提示词
H3 视频提示词(每镜一段)MiniMax H3 的 I2VA 结构:固定对齐指令 + integrated_multimodal_description + overall_soundscape + non_diegetic_music。认领节拍的台词逐字进 <d>[Chinese] …</d>——对白、声景、配乐一段提示词全带上
生成批次单同场景 + 同光照的镜头归一批,共用同一张环境参考图——AI 版的顺场表,脚本自动汇总
配音对齐单每句台词对到镜号——TTS 音频贴到哪一段视频,脚本自动汇总

{baseDir} = 本文件所在目录。脚本 {baseDir}/scripts/novel-storyboard.mjs,零依赖,node 直接跑。

边界(不做的事):不写戏不改台词(novel-script 的活)、不出场景/角色/道具设定图(novel-art / novel-characters 的活)、不做视频生成与剪辑合成。口型/唇形同步暂不管——那是生成管线的事。


Step 0 — 定输入与范围

script.json 是硬前提——分镜离开剧本没有意义,validate/render 都必须给 --script。其余上游按有则用:

  • --outline / --cast:提示词禁人名检查 + 报告里 C01 显示成人名
  • --art:报告里 S01 显示成场景名 + 批次单嵌场景设定图
  • --shots <卡片目录>可选挂载 shot-recipes 的镜头配方卡库(指向 shot-recipes/references/cards,只接受目录不接受导出的 JSON),开第 17 道 shot-recipe 门。没装 shot-recipes 就别给——本 skill 自包含,不依赖它

一次切几集:跟剧本的批次走(剧本写到哪就分到哪),默认一批 ≤ 3 集。

Step 1 — seed 工作底稿

bash
node {baseDir}/scripts/novel-storyboard.mjs seed <script.json> --eps 1-3 > <workdir>/storyboard.json

确定性展开:每场的节拍清单(编号、动作/台词、每拍秒数、说话人)进 seedScenes,这就是切镜时的工作底稿。每拍几秒是算出来的,不要让模型重新估。 shots 留空,切镜才是模型的活。

Step 2 — 逐集分段切镜

每集一份任务,能并发就并发。每份任务拿到:

  • {baseDir}/references/storyboard-pass.md{baseDir}/references/schema.md(读它们,照着做)
  • 该集的 seedScenes 底稿 + 场景卡(art.json 的锚点与光照提示词)+ 角色卡(cast.json 的形象要点)

流程:先按剧情单元分段(每段 9–15 秒、不跨场),段内切 2–5 秒的分镜(对话正反打、关键动作插入特写、进场三件套——切镜语法都在 storyboard-pass.md),每切写一条分镜图提示词。

每段写一条 h3Prompt,照 {baseDir}/references/h3-prompt.md 写(官方方法论的内化版,不依赖任何外部 skill)。官方口径默认英文(promptLang 可切中文),每个镜头独立一行。要点:首行对齐指令和 [Shot k] 切点时刻由分镜秒数推导,一个字符都不许漂(validate 逐字对账);认领台词逐字<d>[Chinese] …</d>;每切的运镜词写进自己那一行;声景与配乐分进后两个字段——声景也是动作指令,画面改了声景一起改

切完把 seedScenes 删掉。

Step 3 — 校验 ⛔ 不能跳

bash
node {baseDir}/scripts/novel-storyboard.mjs validate <storyboard.json> \
  --script <script.json> --outline <outline.json> --cast <cast.json> \
  [--shots </path/to/cards>]

17 道质量门全是代码:节拍全覆盖(分镜级,恰好一次、按顺序、连续)、段 0 < 总秒 ≤ 15、每切 2–5 秒、台词装得进分镜、每集总时长在剧本目标 ±15% 内、同框 ≤ 3 人(超了必须带拆解说明)、段号 E01-01 格式连号、景别短语在分镜图提示词里、风格短语统一style 预设 realistic/ghibli 与角色/场景 skill 同名对齐,同剧分镜图不许画风漂)、运镜用 H3 词表且在自己的 [Shot k] 段落里、H3 对齐指令由分镜结构推导逐字对账 + 切点时刻逐个对认领台词逐字进 <d>提示词语言与 promptLang 一致(双向查:中文写成英文、英文混进中文都拦)、分镜图提示词全英文非空、英文提示词不含角色名(中文 H3 提示词放行)、场次/人物/道具对账剧本、镜头配方对账(可选门,见下)。

有违规逐条修,改完重跑,直到通过。

第 17 道 shot-recipe(可选挂载):给了 --shots 才查,不给就明说跳过。cut 上可以写一个可选的 recipe(配方卡 id,cut 级不是 segment 级多格配方靠连续同 id 的分镜表达,不是数组),门查三条——id 在卡库里、卡片的每条 must_phrases 出现在该切的 frame 里(两边小写化后 includes)、卡片 cuts 下限 ≥ 2 时连续同 id 的分镜数不得低于该下限。卡片的建议景别与运镜不设门,只在报告的「配方」列和 checkup 末尾提示偏离:配方是语汇不是法条,可选挂载的东西一旦变严就没人挂。

Step 4 — 出分镜图(可选)

一切一张 16:9 关键帧,走 codex 内置 $imagegen,读 {baseDir}/references/frame.md 照契约做。要点:

  • 没有 codex 就整步跳过,只交提示词,报告显示占位不装有
  • 参考图是命根子-i 挂上该段场景设定图(该光照状态)+ 画内角色的设定图 + 涉及道具的设定图,提示词只负责取景和此刻的姿态
  • 一格一次调用绝不批量;输出 ./<段号>/f<切序>.png(f1 = 主分镜图,每段一个文件夹)
  • 默认先出第一段的整套分镜图给用户看效果(3–5 张),确认画风和正反打构图再往后补——一集约 30–40 格,错了浪费的是整批
  • 单个失败跳过不阻断,最后汇总说明

Step 5 — 输出与汇报

bash
cd <输出目录>
node {baseDir}/scripts/novel-storyboard.mjs render <剧名>-storyboard.json --md \
  --script <script.json> --outline <outline.json> --art <art.json> > <剧名>-storyboard.md
node {baseDir}/scripts/novel-storyboard.mjs render <剧名>-storyboard.json --html \
  --script <script.json> --outline <outline.json> --art <art.json> > storyboard-report.html

报告界面语言用 --lang zh|en 指定(优先级 --lang > JSON 顶层 lang 字段 > 默认中文)——只切界面标签,与 promptLang(H3 提示词语言)互相独立。render 自动去 images/<镜号>-frame.png 找首帧(批次单还会找场景设定图),先出图再 render。报告含:KPI 带、分镜节奏带(粗分隔 = 段边界、片宽 = 分镜时长占比、颜色深浅 = 景别远近、点击跳段卡)、分集分镜表(主分镜图 + 子分镜条 + 逐切分镜行 + 分镜图/H3 提示词复制按钮)、生成批次单、配音对齐单、质量门、导出 JSON。Markdown 版每段附完整 H3 提示词,直接复制可用。

汇报一句话说清:几集几镜、总时长 vs 目标、几个生成批次、出了几张首帧、报告路径;没过的门和没出的图明说。

最终落地:

<输出目录>/
├── <剧名>-storyboard.json
├── <剧名>-storyboard.md
├── storyboard-report.html         ← 双击就能开
├── manifest.json                  ← export 生成
└── E01-01/                        ← 一段一个文件夹 = 一次 H3 生成的全部材料
    ├── f1.png                     ← 主分镜图(有 codex 才有)
    ├── f2.png …                   ← 子分镜图
    └── prompt.md                  ← H3 提示词(export 生成)

五个 skill 的接力(管线到此闭环)

novel-outline    → outline.json    (什么:结构与分集)
novel-characters → cast.json       (谁:角色设定图)
novel-art        → art.json        (哪里:场景/道具设定图)
novel-script     → script.json     (戏:场次、节拍、台词)
novel-storyboard → storyboard.json (怎么拍:镜头、首帧、批次)

分镜是消费端:seed 吃 script.json,分镜图出图吃 art 和 characters 的设定图当参考,H3 提示词直接下单给视频模型,配音对齐单接 script 台词本的 TTS 产物。五份 JSON 各自的报告都带导出按钮,改完都能喂回各自的 render/validate。

边界

  • 报告界面内置中英(--lang,默认中文);提示词语言由 promptLang 单独控制(默认英文)
  • 秒数是下给视频模型的生成时长不是估算——段上限按你的模型改 params.maxSegmentSeconds,切的节奏区间改 min/maxCutSeconds
  • 口型/唇形同步暂不管——那是生成管线的事
  • 分镜图不追求一次到位——它是给视频模型的构图锚,构图对、资产对就够,微调交给重生成

门失败会累积

validatecheckup 每次都把门的结果追加到当前目录.gates.jsonl;跑 stats 汇总:

bash
node {baseDir}/scripts/novel-storyboard.mjs stats

回答三件事:哪道门最常响(那条规则模型最常无视,该改的是措辞)、哪道门从没响过(可能是死门,也可能规则已被内化)、失败详情长什么样(反复出现却没有门的那类问题,只能靠人看)。

不想记加 --no-log;写不进去静默跳过,不影响校验。

自测

bash
node {baseDir}/scripts/selftest.mjs

254 项断言,不调模型、不花额度。17 道质量门每一道都有击穿用例。改完脚本先跑这个。

自带样例

{baseDir}/examples/渡口-storyboard.json:《渡口》第 1 集完整分镜——10 段 34 切认领剧本全部 35 拍,平均 3.5 秒一切,共 119 秒 / 目标 120 秒,2 个生成批次,每段带完整的 H3 视频提示词(多图对齐 + 切点时刻全部对账通过)。当质量基准,也是自测夹具。

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Novel Storyboard AI skill do?

给 AI 短剧出分镜:三层结构——段(一次视频生成,≤15 秒)→ 分镜(段内 2–5 秒的剪切,认领剧本节拍) → 分镜图(每切一张关键帧:主分镜图钉 0.00 秒,子分镜图钉各自切点)。 每段自带一条 MiniMax H3 视频提示词(官方口径默认英文、逐镜换行,promptLang 可切中文):对齐指令和 [Shot k] 切点时刻由分镜结构推导、逐字对账,台词逐字进 <d> 块(写法规范已内化为 references/h3-prompt.md,不依赖外部 skill)。 产出 storyboard.json + Markdown + 单页评审报告(分镜节奏带 / 分集分镜表 / 生成批次单 / 配音对齐单,含导出 JSON)。分镜图出图拿场景与角色设定图当参考图走 codex $imagegen(可选)。 17 道质量门全部由脚本确定性检查(第 17 道 shot-recipe 可选:挂上 shot-recipes 卡库才查,不挂就明说跳过); export 一键导出 H3 投产包(每段提示词 + 按 Picture 序的分镜图清单)。零依赖、零 API key,用当前会话额度。 Use when asked to 分镜、出分镜、镜头表、切镜、storyboard for AI short drama。

Why use Novel Storyboard on TypingMind?

Because you install it once and use it with any model. Novel Storyboard is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Novel Storyboard in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/eternityspring/shuohao-skills/tree/main/skills/novel-storyboard. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Novel Storyboard?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Novel Storyboard?

As many as you like. As long as a model supports skills, you can use Novel Storyboard with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Novel Storyboard AI skill free?

Yes. It is published on GitHub by eternityspring under the Apache-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇