Google Sheets MCP
Enables AI agents to interact directly with Google Sheets to create, read, and edit spreadsheets through the Model Context Protocol. It supports a wide range of actions including cell manipulation, row and column management, and sheet organization.
README
Google Sheets MCP
A Model Context Protocol (MCP) connector for Google Sheets that allows AI agents to interact with spreadsheets directly.
Demo
https://github.com/user-attachments/assets/cc4729d9-4e6e-437b-848b-6da9a09418c3
Setup
- Clone this repository:
git clone https://github.com/mkummer225/google-sheets-mcp
cd google-sheets-mcp
-
Install dependencies:
npm install -
Build:
npm run build -
Create OAuth credentials in Google Cloud Platform:
- Create a new project in Google Cloud Console
- Enable the Google Sheets API
- Configure the OAuth consent screen
- Create OAuth client ID credentials (Desktop application) with an appropriate redirect URI (ex: http://localhost:3000/oauth2callback)
- Download the credentials and save as
gcp-oauth.keys.jsonin thedistsubdirectory
-
Start the MCP server (you'll automatically be prompted to authenticate/re-authenticate your Google account when necessary):
npm run start
Hosted / Railway (SSE transport)
If you need a URL-based MCP endpoint (e.g., for the Agent Toolkit’s “Connect to MCP Server” flow), run with the SSE transport and deploy to a host like Railway:
- Build:
npm run build - Set env vars (do not commit secrets):
MCP_TRANSPORT=ssePORT=3000(or your platform’s port)GSHEETS_OAUTH_JSON— your OAuth client JSON (plain or base64-encoded). Required if the filedist/gcp-oauth.keys.jsonis not present.GSHEETS_CREDENTIALS_JSON— a pre-authorized credentials JSON (plain or base64-encoded). Use this to avoid interactive auth on hosted environments.
- Deploy
dist/to Railway and runnode dist/index.js. - In your client/UI, point the MCP URL to
https://<your-railway-app>.up.railway.app/sse(the server will direct POSTs to/messagesvia the SSE endpoint event).
Note: Keep *.keys.json and credential files out of git. Use env vars or your host’s secrets manager instead.
Usage
Sample config:
{
"mcpServers": {
"google-sheets-mcp": {
"command": "node",
"args": [
"/{path_to_dir}/google-sheets-mcp/dist/index.js"
]
}
}
}
Then you should be able to simply specify your spreadsheetId or ask your agent to create a new one for you.
Available Actions
| Action | Description |
|---|---|
refresh_auth |
Re-authenticate your Google Account when credentials expire |
list_sheets |
List all sheets/tabs in a Google Spreadsheet |
create_sheet |
Create a new sheet/tab in a Google Spreadsheet |
create_spreadsheet |
Create a new Google Spreadsheet |
read_all_from_sheet |
Read all data from a specified sheet |
read_headings |
Read the column headings from a sheet |
read_rows |
Read specific rows from a sheet |
read_columns |
Read specific columns from a sheet |
edit_cell |
Edit a single cell in a sheet |
edit_row |
Edit an entire row in a sheet |
edit_column |
Edit an entire column in a sheet |
insert_row |
Insert a new row at specified position |
insert_column |
Insert a new column at specified position |
rename_sheet |
Rename a sheet/tab in a spreadsheet |
rename_doc |
Rename a Google Spreadsheet |
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。