google-sheets-mcp
Local Model Context Protocol server for reading and writing Google Sheets through Google Sheets API v4.
README
google-sheets-mcp
English | Korean
Local Model Context Protocol server for reading and writing Google Sheets through Google Sheets API v4.
This project is designed for local Codex/AI-agent workflows. It supports both service account credentials and OAuth user credentials.
Features
- Read spreadsheet metadata and worksheet tabs.
- Read values from A1 ranges.
- Update exact ranges.
- Append rows.
- Batch update multiple ranges.
- Clear values while preserving formatting.
- Report authentication status without exposing private keys or tokens.
MCP Tools
auth_statusget_spreadsheetlist_worksheetsread_rangeupdate_rangeappend_rowsbatch_update_valuesclear_range
Install
npm.cmd install
npm.cmd run build
On macOS/Linux, use npm instead of npm.cmd.
Authentication: Service Account
Recommended for automation.
- Create or choose a Google Cloud project.
- Enable Google Sheets API.
- Create a service account and download a JSON key.
- Put the JSON key at:
secrets/google-service-account.json
- Share each target spreadsheet with the service account email from the JSON key:
{
"client_email": "your-service-account@your-project.iam.gserviceaccount.com"
}
Give Viewer access for read-only use or Editor access for writes.
The secrets/ directory is committed with only .gitkeep. All credential files inside it are ignored.
Authentication: Custom Credential Path
If you do not want credentials under the repository directory, create a local .env file:
GOOGLE_APPLICATION_CREDENTIALS=C:/secure/path/google-service-account.json
or provide inline JSON through a secret manager/environment variable:
GOOGLE_SERVICE_ACCOUNT_JSON={"type":"service_account",...}
Credential lookup order:
GOOGLE_APPLICATION_CREDENTIALSGOOGLE_SERVICE_ACCOUNT_JSON./secrets/google-service-account.json
Authentication: OAuth User Account
Use OAuth when the MCP should act as your Google user account.
- Create OAuth Client credentials in Google Cloud.
- Add this redirect URI:
http://127.0.0.1:3333/oauth2callback
- Create
.env:
GOOGLE_OAUTH_CLIENT_ID=your-client-id.apps.googleusercontent.com
GOOGLE_OAUTH_CLIENT_SECRET=your-client-secret
OAUTH_CALLBACK_PORT=3333
- Generate a refresh token:
npm.cmd run oauth:init
- Open the printed URL, approve access, and add the printed refresh token to
.env:
GOOGLE_OAUTH_REFRESH_TOKEN=your-refresh-token
Register With Codex
Build first:
npm.cmd run build
Then register the MCP server:
codex mcp add google-sheets-local -- node <path-to-google-sheets-mcp>/dist/index.js
Recommended ~/.codex/config.toml entry:
[mcp_servers.google-sheets-local]
command = "node"
args = ["<path-to-google-sheets-mcp>/dist/index.js"]
cwd = "<path-to-google-sheets-mcp>"
Restart Codex after changing MCP configuration or credentials.
Usage Notes
- Use
auth_statusfirst to verify the active authentication mode. - Use
list_worksheetsbefore reading or writing ranges. - For destructive operations, prefer reading the target range first.
- Use
batch_update_valuesfor large table edits to reduce Google API request count.
Security
Never commit credentials. The following are ignored by default:
.env.env.*except.env.examplecredentials/secrets/*exceptsecrets/.gitkeep*.service-account.json*.credentials.json*.token.json
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。
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。
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。
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。