@datalanguage/datagraphs-mcp-server
MCP server that connects AI assistants to the Data Graphs knowledge graph platform, enabling natural language search, exploration, and querying of graph data.
README
@datalanguage/datagraphs-mcp-server
An MCP (Model Context Protocol) server that exposes the Data Graphs knowledge graph platform to AI assistants such as Claude, ChatGPT, Cursor, and other MCP-compatible clients.
Prerequisites
- Node.js 18+
- A Data Graphs project with an API key
Installation
npm install @datalanguage/datagraphs-mcp-server
Or run directly with npx:
npx @datalanguage/datagraphs-mcp-server
Configuration
Obtaining Credentials
To generate the API key, client ID, and client secret:
- Sign in to Data Graphs and open your project.
- Click Manage Project.
- Go to Applications and click Add Application.
- Once the application is created, the API key, client ID, and client secret will be displayed.
Important: The client secret is shown only once at creation time. Copy it immediately and store it somewhere secure — if lost, you will need to rotate the credentials or create a new application to obtain a new secret.
Your project ID can also be found on the project page in Data Graphs.
Environment Variables
The server reads credentials from environment variables:
| Variable | Required | Description |
|---|---|---|
DATAGRAPHS_PROJECT_ID |
Yes | Your Data Graphs project ID |
DATAGRAPHS_API_KEY |
Yes | API key for authentication |
DATAGRAPHS_CLIENT_ID |
For GQL | OAuth client ID (required for graph_search) |
DATAGRAPHS_CLIENT_SECRET |
For GQL | OAuth client secret (required for graph_search) |
MCP Client Setup
Claude Desktop
Add to your claude_desktop_config.json:
{
"mcpServers": {
"datagraphs": {
"command": "npx",
"args": ["-y", "@datalanguage/datagraphs-mcp-server"],
"env": {
"DATAGRAPHS_PROJECT_ID": "your-project-id",
"DATAGRAPHS_API_KEY": "your-api-key"
}
}
}
}
Claude Code
claude mcp add datagraphs -- npx -y @datalanguage/datagraphs-mcp-server
Then set the required environment variables in your shell before launching Claude Code.
Cursor
Add to your Cursor MCP settings (.cursor/mcp.json):
{
"mcpServers": {
"datagraphs": {
"command": "npx",
"args": ["-y", "@datalanguage/datagraphs-mcp-server"],
"env": {
"DATAGRAPHS_PROJECT_ID": "your-project-id",
"DATAGRAPHS_API_KEY": "your-api-key"
}
}
}
}
Tools
The server provides the following tools:
list_datasets
List all datasets available in the Data Graphs project.
| Parameter | Type | Description |
|---|---|---|
pageSize |
number | Number of results per page |
pageNo |
number | Page number to return |
search_concepts
Search for concepts (nodes) across datasets using text queries, filters, and facets.
| Parameter | Type | Description |
|---|---|---|
dataset |
string | Dataset ID to search (defaults to all datasets) |
q |
string | Text search query |
filter |
string | NQL filter expression |
facets |
string[] | Property names to return facet counts for |
pageSize |
number | Number of results per page (default 10) |
pageNo |
number | Page number to return (default 1) |
sort |
string | Sort order, e.g. name:asc,modified:desc |
embed |
string[] | Relationship properties to embed (use ["_all"] to embed all) |
fields |
string[] | Specific fields to return for each concept |
includeDateFields |
boolean | Include created and last modified dates |
get_concept
Get a single concept (node) by its URN identifier.
| Parameter | Type | Description |
|---|---|---|
id |
string | Concept URN, e.g. urn:project:Type:identifier |
embed |
string[] | Relationship properties to embed (use ["_all"] to embed all) |
fields |
string[] | Specific fields to return |
includeDateFields |
boolean | Include created and last modified dates |
graph_search
Execute a GQL graph query against the knowledge graph. Requires DATAGRAPHS_CLIENT_ID and DATAGRAPHS_CLIENT_SECRET.
| Parameter | Type | Description |
|---|---|---|
query |
string | GQL query, e.g. MATCH (n:Person) RETURN n LIMIT 10 |
search_transactions
Search the transaction/audit log for the project.
| Parameter | Type | Description |
|---|---|---|
filter |
string | NQL filter expression |
pageSize |
number | Number of results per page |
pageNo |
number | Page number to return |
sort |
string | Sort order, e.g. property:asc |
get_model
Get the currently active data model for the project. Takes no parameters.
Development
git clone <repo-url>
cd datagraphs-client-mcp-server
npm install
npm start
License
UNLICENSED
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。