nodeland-mcp
MCP server that lets agents create and update persistent concept maps via the NodeLand API, returning shareable URLs for humans to view, embed, and keep editing.
README
nodeland-mcp
NodeLand is the persistent destination where agent output becomes an editable map — not a one-shot Mermaid render that dies with the chat.
This MCP server talks to the NodeLand Maps API. Agents emit concepts + relations; NodeLand stores the map, runs layout, and returns a URL the human can open, share, embed, and keep editing.
Demo maps (open these)
Replace these with maps you create during a real agent session before publishing:
- Example concept map
- Create your own with Claude/Cursor below, then paste the live URL here — the artifact is the demo.
<!-- Record 3 short GIFs (agent prompt → tool call → map in browser) into docs/ before npm publish:
-->
What you get
| Tool | When to use |
|---|---|
create_map |
Turn a topic into a shareable map (returns url + embedUrl) |
add_to_map |
Keep writing into the same map across a conversation |
get_map |
Read concepts/relations summary (not raw Cytoscape JSON) |
list_maps |
Find an existing map id |
Resource: nodeland://templates — named layouts (StandardMindMapTemplate, TonyBuzzanTemplate, Tree TD / Tree LR, …).
Maps are public by default so the link works without login. Pass isPublic: false only when the user asks for private.
1. Get an API key (the real onboarding step)
- Sign up / log in at nodeland.io
- Open Settings → API Keys
- Generate a key with
maps:readandmaps:write - Copy it once — it is shown only at creation time
2. Install in your agent
Claude Desktop
Add to claude_desktop_config.json:
{
"mcpServers": {
"nodeland": {
"command": "npx",
"args": ["-y", "nodeland-mcp"],
"env": {
"NODELAND_API_KEY": "YOUR_API_KEY"
}
}
}
}
Claude Code
claude mcp add nodeland --env NODELAND_API_KEY=YOUR_API_KEY -- npx -y nodeland-mcp
Cursor
Cursor Settings → MCP → add server:
{
"mcpServers": {
"nodeland": {
"command": "npx",
"args": ["-y", "nodeland-mcp"],
"env": {
"NODELAND_API_KEY": "YOUR_API_KEY"
}
}
}
}
Optional staging:
"NODELAND_BASE_URL": "https://staging.nodeland.io"
3. Try it
Ask the agent:
Map the central concepts of [topic], put a short markdown description on each node, and give me the NodeLand link.
Then open the URL in a private/incognito window — it should render without signing in.
Config
| Env | Required | Default |
|---|---|---|
NODELAND_API_KEY |
yes | — |
NODELAND_BASE_URL |
no | https://nodeland.io |
The key is never logged. Missing key fails fast with a readable message.
Local development
git clone https://github.com/nodeland/nodeland-mcp.git
cd nodeland-mcp
npm install
npm run build
NODELAND_API_KEY=... npm start
Smoke test (creates a public map and fetches the URL without credentials):
NODELAND_API_KEY=... NODELAND_BASE_URL=https://staging.nodeland.io npm run smoke
Inspector:
npx @modelcontextprotocol/inspector npx nodeland-mcp
License
MIT
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。