Timeverse InfCanvas MCP Server
Enables AI clients to read and write local infinite canvas data via MCP protocol, with support for image generation and web visualization.
README
Timeverse InfCanvas MCP Server
本地无限画布的 MCP Server,让 AI 客户端(Claude Desktop / Cursor 等)通过 Model Context Protocol 直接读写本地画布数据。
项目结构
timeverse-InfCanvas-mcp/
├── mcp-server/ ← MCP 协议服务(核心)
│ ├── server.mjs MCP Server 主入口
│ ├── canvas-store.mjs 画布数据读写
│ ├── image-generator.mjs 大模型生图
│ ├── fractional-indexing.mjs z-order 排序
│ ├── overlap-detector.mjs 碰撞检测
│ ├── seed-demo.mjs 演示数据生成
│ ├── test.mjs 冒烟测试
│ ├── package.json 依赖声明
│ └── node_modules/
├── web-server/ ← HTTP 可视化服务
│ └── web-server.mjs 画布浏览器查看 + 编辑 API
├── client/ ← React + tldraw 编辑器
│ ├── src/ React 源码
│ ├── dist/ 构建产物
│ └── package.json
├── canvas/ ← 画布数据目录
├── package.json 根项目脚本
├── LICENSE
└── README.md
架构
AI 客户端 ──(stdio JSON-RPC)──> mcp-server/server.mjs
│
├── canvas-store.mjs canvas/*.json
├── image-generator.mjs 模型 API
├── fractional-indexing.mjs
└── overlap-detector.mjs
浏览器 ──(HTTP)──> web-server/web-server.mjs
│
├── canvas/*.json (读取画布)
└── client/dist/ (前端静态文件)
三个模块,各司其职:
| 模块 | 职责 | 技术栈 |
|---|---|---|
| mcp-server | MCP 协议通信,AI 调用工具读写画布 | Node.js ESM, @modelcontextprotocol/sdk |
| web-server | HTTP 可视化,浏览器实时查看/编辑画布 | Node.js 原生 http 模块(零依赖) |
| client | 前端编辑器(可选),tldraw 专业图形界面 | React 18, tldraw 5, Vite |
安装
# 克隆项目
git clone <repo-url>
cd timeverse-InfCanvas-mcp
# 安装 MCP Server 依赖
cd mcp-server && npm install
# (可选)构建前端编辑器
cd ../client && npm install && npm run build
在 AI 客户端中配置
MCP Server 通过 stdio 协议与 AI 客户端通信,配置时指向 mcp-server/server.mjs。
Claude Desktop
编辑 claude_desktop_config.json:
{
"mcpServers": {
"infcanvas": {
"command": "node",
"args": ["D:/path/to/timeverse-InfCanvas-mcp/mcp-server/server.mjs"],
"env": {
"INF_CANVAS_DIR": "D:/path/to/your-project",
"REPLICATE_API_KEY": "r-xxxxx"
}
}
}
}
Cursor
编辑 .cursor/mcp.json:
{
"mcpServers": {
"infcanvas": {
"command": "node",
"args": ["D:/path/to/timeverse-InfCanvas-mcp/mcp-server/server.mjs"]
}
}
}
自研客户端
支持以下方式集成:
方式一:JSON 配置(适用于支持 MCP 配置文件的客户端)
{
"mcpServers": {
"infcanvas": {
"command": "node",
"args": ["D:/path/to/timeverse-InfCanvas-mcp/mcp-server/server.mjs"],
"env": {
"INF_CANVAS_DIR": "D:/my-project"
}
}
}
}
方式二:通过 npx(无需本地安装,自动从 npm 拉取)
{
"mcpServers": {
"infcanvas": {
"command": "npx",
"args": ["-y", "timeverse-infcanvas-mcp"],
"env": {
"INF_CANVAS_DIR": "D:/my-project"
}
}
}
}
注意:
timeverse-infcanvas-mcp尚未发布到 npm,发布后方可使用此方式。
方式三:代码中通过 stdio 启动子进程
const { spawn } = require('node:child_process')
const cp = spawn('node', ['D:/path/to/mcp-server/server.mjs'], {
env: { ...process.env, INF_CANVAS_DIR: 'D:/my-project' },
stdio: ['pipe', 'pipe', 'pipe'],
})
// 通过 stdin/stdout 收发 JSON-RPC 2.0 消息
暴露的 MCP 工具
所有数据操作工具都需要 projectDir 参数(指定项目目录,在其下创建 canvas/ 文件夹):
| 工具 | 用途 | 必填参数 |
|---|---|---|
get_canvas_snapshot |
获取画布全量数据 | projectDir |
get_selection |
读取当前用户选区 | projectDir |
list_pages |
列出所有页面 | projectDir |
create_page |
新建页面 | projectDir |
switch_page |
切换当前页面 | projectDir, pageId |
create_ai_image_holder |
创建 AI 图片占位框 | projectDir |
insert_image |
插入图片(本地路径或 URL) | projectDir, sourcePath |
fill_ai_holder |
把已有 asset 填入占位框 | projectDir, holderId, assetId |
update_shape |
更新 shape 位置/属性 | projectDir, shapeId |
update_shape_status |
流式更新生成状态与进度 | projectDir, shapeId, status |
delete_shapes |
删除多个 shape | projectDir, shapeIds |
generate_image |
大模型生图并直接填入画布 | projectDir, prompt |
start_web_server |
启动 HTTP 可视化服务(无 projectDir 也可调用) | — |
start_web_server
AI 调用此工具后,会在本地启动 HTTP 服务,打开浏览器即可实时查看/操作画布:
{
"name": "start_web_server",
"arguments": { "port": 3000 }
}
返回 { "url": "http://localhost:3000", "status": "started" }。重复调用不会重复启动,浏览器打开 http://localhost:3000 即可查看画布。
手动启动
# MCP Server(stdio 交互,通常由 AI 客户端自动拉起)
npm start
# Web 可视化服务
npm run web
# 浏览器打开 http://localhost:3000
# 前端开发模式
npm run dev:client
# 构建前端
npm run build:client
环境变量
| 变量 | 用途 | 必填 |
|---|---|---|
INF_CANVAS_DIR |
兜底项目目录(不填则每次调用传 projectDir) |
否 |
OPENAI_API_KEY |
OpenAI API Key(DALL-E 3 生图) | 按需 |
OPENAI_API_ENDPOINT |
OpenAI 端点,默认 https://api.openai.com/v1 |
否 |
REPLICATE_API_KEY |
Replicate API Token(Flux / SDXL 生图) | 按需 |
CUSTOM_API_ENDPOINT |
自研模型 HTTP 端点 | 按需 |
CUSTOM_API_KEY |
自研模型 API Key | 按需 |
大模型生图
支持的模型(通过环境变量路由):
| model | 依赖 |
|---|---|
dall-e-3 |
OPENAI_API_KEY |
flux-schnell |
REPLICATE_API_KEY |
sdxl |
REPLICATE_API_KEY |
custom |
CUSTOM_API_ENDPOINT + CUSTOM_API_KEY |
工具参数中传 endpoint 和 apiKey 可临时覆盖环境变量。
测试
cd mcp-server
node test.mjs
冒烟测试会验证 MCP 协议通信、数据隔离、工具调用等核心功能。
画布数据格式
<project>/canvas/
├── canvas.json 画布索引(页面、shape 列表、资源)
├── selection.json 当前选区
├── view-state.json 视口状态(相机位置、当前页)
└── shapes/ 独立 shape 文件(每个 .json)
三方鸣谢
- tldraw — 前端编辑器基于 tldraw v5,提供专业级的无限画布交互体验
- Model Context Protocol SDK — MCP 协议实现
- React + Vite — 前端工程化
- Replicate — Flux / SDXL 模型推理
- OpenAI — DALL-E 3 图片生成
许可证
MIT © 2024-2026 TimeVerse Studio
完整许可见 LICENSE。
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.