Novel Writing Assistant MCP Server
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Novel Writing Assistant MCP Server 创建一个新章节,标题'初入江湖',大纲:主角离开家乡"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
小说写作辅助 MCP 服务器
这是一个专门为小说创作设计的 MCP (Model Context Protocol) 服务器,帮助你管理人物、章节、世界观和剧情线索。
功能特性
1. 人物管理
创建和管理小说人物
记录人物性格、背景、外貌、关系
快速查询人物信息
2. 章节管理
按章节组织小说内容
记录章节大纲和正文
追踪每章涉及的人物
3. 世界观设定
保存各类世界观设定(魔法体系、地理、历史等)
分类管理不同类型的设定
随时查阅和更新
4. 剧情线索
追踪多条剧情线的发展
标记剧情状态(进行中/已完结/伏笔)
关联相关章节
5. 灵感扩展
提供创作建议和灵感
支持人物发展、情节转折、冲突设计、场景描写等多个维度
Related MCP server: novel-mcp
安装步骤
确保已安装 Node.js (版本 18 或更高)
进入项目目录并安装依赖:
cd xiaoshuo-mcp
npm install配置 Kiro MCP
在 .kiro/settings/mcp.json 中添加:
{
"mcpServers": {
"xiaoshuo": {
"command": "node",
"args": ["你的路径/xiaoshuo-mcp/index.js"],
"disabled": false
}
}
}使用示例
创建人物
使用 chuangJianRenWu 工具:
- mingZi: "张三"
- xingGe: "勇敢、正直、有时冲动"
- beiJing: "出身武林世家,自幼习武"
- waiMao: "身材高大,剑眉星目"创建章节
使用 chuangJianZhangJie 工具:
- zhangJieHao: 1
- biaoTi: "初入江湖"
- daGang: "主角离开家乡,第一次踏入江湖"
- neiRong: "章节正文内容..."
- sheJiRenWu: ["张三", "李四"]设置世界观
使用 sheZhiShiJieGuan 工具:
- leiXing: "魔法体系"
- neiRong: "本世界的魔法分为五大元素..."添加剧情线
使用 tianJiaJuQingXian 工具:
- mingCheng: "寻找神器"
- miaoShu: "主角寻找传说中的神器以对抗邪恶势力"
- zhuangTai: "进行中"
- xiangGuanZhangJie: [1, 3, 5]获取创作灵感
使用 kuoZhanLingGan 工具:
- leiXing: "情节转折"
- shangXiaWen: "主角刚刚获得了一个重要线索"数据存储
所有数据保存在 xiaoshuo-mcp/xiaoshuo-shuju/ 目录下:
renwu/- 人物信息zhangjie/- 章节内容shijieguan/- 世界观设定juqing/- 剧情线索
数据以 JSON 格式存储,方便备份和迁移。
可用工具列表
chuangJianRenWu- 创建新人物huoQuRenWu- 获取人物信息lieJuRenWu- 列出所有人物chuangJianZhangJie- 创建新章节huoQuZhangJie- 获取章节内容lieJuZhangJie- 列出所有章节sheZhiShiJieGuan- 设置世界观huoQuShiJieGuan- 获取世界观设定tianJiaJuQingXian- 添加剧情线huoQuJuQingXian- 获取所有剧情线kuoZhanLingGan- 获取创作灵感
注意事项
所有数据本地存储,注意定期备份
人物名称和章节号作为唯一标识,不要重复
建议按顺序创建章节,便于管理
⚠️ 重要:写作规范
在使用AI写作前,请务必阅读 xiaoshuo-shuju/写作规范.md
核心要求:
写新章节前必须先读取前3章内容,确保逻辑连贯
严格遵守800章大纲,不要跳跃或提前写后面的情节
去除AI味:避免"深吸一口气"、"眼中闪过"等套话
对话生活化:用口语、语气词,不同人物说话风格不同
加入生活细节:具体的食物、天气、小动作等
逻辑一致性:时间、地点、人物状态、身份要前后一致
详细规范请查看: xiaoshuo-shuju/写作规范.md
祝你创作愉快!📖✨
Available Tools
11 toolschuangJianRenWuC
创建新的小说人物,包含姓名、性格、背景等信息
Inputs: mingZi (required), guanXi, waiMao, xingGe, +1 more.
chuangJianZhangJieA
创建新章节,记录章节内容和大纲。
⚠️ 重要写作规范(必须遵守):
写新章节前必须先用huoQuZhangJie读取前3章内容,确保逻辑连贯
严格遵守800章大纲(xiaoshuo-shuju/800章大纲.md),不要跳跃或提前写后面的情节
去除AI味:避免"深吸一口气"、"眼中闪过一丝"、"嘴角勾起一抹"等套话
对话生活化:用口语、语气词(嗯、啊、哈哈),不同人物说话风格要不同
加入生活细节:具体的食物、天气、小动作、现代元素(手机、电脑)
逻辑一致性:
时间地点要与上一章衔接
人物要记得之前发生的事
唐双是明月会真正老大,老鬼阿豹都知道,不要重复宣布身份
已展示的能力要有来源铺垫(师傅教的、古籍学的)
情绪真实:不要每次都"震惊"、"惊讶",用更细腻的描写
节奏把控:每段都要推进剧情或刻画人物,不要水字数
详细规范见: xiaoshuo-shuju/写作规范.md
Inputs: biaoTi (required), zhangJieHao (required), daGang, neiRong, +1 more.
huoQuJuQingXianB
获取所有剧情线索
Inputs: no named parameters.
huoQuRenWuC
获取指定人物的详细信息
Inputs: mingZi (required).
huoQuShiJieGuanC
获取世界观设定
Inputs: leiXing.
huoQuZhangJieB
获取指定章节的内容
Inputs: zhangJieHao (required).
kuoZhanLingGanC
基于现有内容生成创作灵感和建议
Inputs: leiXing (required), shangXiaWen.
lieJuRenWuB
列出所有已创建的人物
Inputs: no named parameters.
lieJuZhangJieB
列出所有章节概览
Inputs: no named parameters.
sheZhiShiJieGuanC
设置或更新世界观设定
Inputs: leiXing (required), neiRong (required).
tianJiaJuQingXianC
添加剧情线索,追踪故事发展
Inputs: miaoShu (required), mingCheng (required), zhuangTai, xiangGuanZhangJie.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 11 tools
The set reuses the same tool name 'huoQuZhangJie' for both '获取所有章节概览' and '获取指定章节的内容', and it appears duplicated. An agent cannot reliably distinguish which tool to call from the name alone, making selection highly ambiguous.
Most names follow a readable pinyin verb_noun pattern (huoQu..., chuangJian..., lieJu...). However, assigning the same name 'huoQuZhangJie' to different operations breaks the predictability expected from a consistent naming scheme.
With 7 tools, the count is within a reasonable range for a focused writing assistant. It is not excessive, though the duplicate name and missing related operations make the effective surface feel smaller than claimed.
The surface covers basic create/list for characters and create/get for chapters, but lacks update/delete operations for characters and chapters, outline management despite referencing an 800-chapter outline, search, and consistency-help tools. These are significant lifecycle gaps that would force workarounds.
Related MCP Connectors
Story continuity, novel memory, and series bible for fiction writers, shared with AI assistants
Writing studio for novels and screenplays: read your projects and run cited fact-checks.
Write a novel in the Claude or ChatGPT you already pay for. Keeps your book, cast and timeline.
A structured story bible for long-form fiction: scenes, characters, world, plot and foreshadowing.
Related MCP Servers
- -licenseNot gradedqualityNot gradedmaintenanceA novel management tool that uses SQLite database to store and manage novel information including chapters, characters, and plot outlines. Provides database operations and SQL query capabilities for writers to organize their creative work through natural language.1,107 npm-
- AlicenseAqualityCmaintenanceEnables AI tools to collaboratively write novels by managing chapters, characters, and story state through commands like validate, context, draft, review, and approve.8MIT
- AlicenseBqualityBmaintenanceEnables writers and AI agents to preserve continuity in long-form fiction by maintaining a narrative knowledge graph and exposing MCP tools for querying outlines, entities, references, and consistency diagnostics.14MIT
- AlicenseNot gradedqualityBmaintenanceEnables AI-driven long-form novel creation and management, including chapter generation, character and timeline tracking, semantic memory retrieval, version savepoints, and deep consistency checking across multiple projects through MCP tools, with support for local LM Studio or any OpenAI-compatible API.16MIT