1Password MCP Server

1Password MCP Server

Connects MCP-compatible AI clients to 1Password vaults via a Service Account, enabling password management and vault operations through natural language.

Category
访问服务器

README

1Password MCP Server

CI npm License

A community-built Model Context Protocol (MCP) server that connects MCP-compatible AI clients (Claude Desktop, VS Code Copilot, OpenAI Codex, Gemini, etc.) to 1Password vaults via a Service Account.

Not an official 1Password product. This is a community project.


Features

Tools (8)

Tool Description
vault_list List all accessible vaults
item_lookup Search items by title in a vault
item_delete Delete an item from a vault
password_create Create a new password/login item
password_read Retrieve a password via secret reference (op://vault/item/field) or vault/item ID
password_update Rotate/update an existing password
password_generate Generate a cryptographically secure random password
password_generate_memorable Generate a memorable passphrase from ~500 dictionary words

Prompts (4)

Prompt Description
generate-secure-password Guided workflow to generate and store a secure password
credential-rotation Step-by-step credential rotation: read, generate, update, verify
vault-audit Audit vault contents: list items, categorize, flag concerns
secret-reference-helper Construct op://vault/item/field references interactively

Resources (3)

Resource URI Description
1password://config Current server configuration (non-secret)
1password://vaults Browsable list of all accessible vaults
1password://vaults/{vaultId}/items Browsable list of items in a vault

Quick Start

Prerequisites

Claude Desktop / VS Code / IDEs (JSON)

{
  "mcpServers": {
    "1password": {
      "command": "npx",
      "args": ["-y", "@takescake/1password-mcp"],
      "env": {
        "OP_SERVICE_ACCOUNT_TOKEN": "YOUR_SERVICE_ACCOUNT_TOKEN"
      }
    }
  }
}

macOS Keychain (JSON)

If you do not want to store the service account token directly in your MCP config, macOS users can store it in Keychain and configure the server to read it at startup instead:

{
  "mcpServers": {
    "1password": {
      "command": "npx",
      "args": ["-y", "@takescake/1password-mcp"],
      "env": {
        "OP_KEYCHAIN_SERVICE": "op-service-account-claude-automation",
        "OP_KEYCHAIN_ACCOUNT": "your-macos-username"
      }
    }
  }
}

Precedence is: CLI arguments (--service-account-token / --token) > OP_SERVICE_ACCOUNT_TOKEN > macOS Keychain lookup. OP_KEYCHAIN_ACCOUNT is optional if your Keychain service name is already unique enough.

OpenAI Codex (TOML)

Option A (stores the token in config):

[mcp_servers."1password"]
command = "npx"
args = ["-y", "@takescake/1password-mcp"]

[mcp_servers."1password".env]
OP_SERVICE_ACCOUNT_TOKEN = "YOUR_SERVICE_ACCOUNT_TOKEN"

Option B (recommended: does NOT store the token in Codex config):

[mcp_servers."1password"]
command = "npx"
args = ["-y", "@takescake/1password-mcp"]
env_vars = ["OP_SERVICE_ACCOUNT_TOKEN"]

Then set OP_SERVICE_ACCOUNT_TOKEN in your shell/session/CI environment.

Note: codex mcp add ... --env OP_SERVICE_ACCOUNT_TOKEN=... writes the token into Codex config. Use env_vars if you want the config to reference only the variable name.

On macOS, you can also omit OP_SERVICE_ACCOUNT_TOKEN and set OP_KEYCHAIN_SERVICE (plus optional OP_KEYCHAIN_ACCOUNT) to read the token from Keychain at startup.

CLI Options

--service-account-token <token>   1Password service account token
--log-level <level>               Log level: error, warn, info, debug (default: info)
--integration-name <name>         Custom integration name for 1Password SDK
--integration-version <version>   Custom integration version

Security & Privacy

Read this before using.

  • LLM privacy risk -- Secrets retrieved/created may be sent to your LLM provider and could be retained depending on your provider/account settings.
  • No E2E encryption in MCP -- Secrets are plaintext inside the MCP workflow and in transit to the model. They are encrypted only once stored in 1Password.
  • Intended use -- Best for automated/disposable credentials (dev DB creds, bot/service accounts, CI tokens).
  • Avoid high-stakes secrets -- Do not use for banking, primary personal accounts, or other sensitive credentials. Use dedicated automation vaults.
  • Token security -- Treat the Service Account Token like a master key. Rotate immediately if exposed.
  • Config files -- Keep MCP config files out of version control (add to .gitignore).
  • Secret references -- Prefer op://... references over copying raw passwords into prompts or files.
  • Least privilege -- Use dedicated vaults and limited-scope service accounts for automation workflows.

Development

# Clone and install
git clone https://github.com/CakeRepository/1Password-MCP.git
cd 1Password-MCP
npm install

# Build
npm run build

# Run tests
npm test

# Type-check
npm run lint

# Watch mode (dev)
npm run dev

Project Structure

src/
  index.ts              # Server entrypoint
  types.ts              # Shared type definitions
  logger.ts             # Structured logging (stderr)
  config.ts             # CLI args, env vars, constants
  client.ts             # 1Password SDK client singleton
  utils.ts              # Result helpers, password generation
  tools/                # MCP tool handlers
    index.ts
    vault-list.ts
    item-lookup.ts
    item-delete.ts
    password-create.ts
    password-read.ts
    password-update.ts
    password-generate.ts
    password-generate-memorable.ts
  prompts/              # MCP prompt definitions
    index.ts
  resources/            # MCP resource definitions
    index.ts

See CONTRIBUTING.md for contribution guidelines.


License

Apache License 2.0

推荐服务器

Baidu Map

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。

官方
精选
JavaScript
Playwright MCP Server

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。

官方
精选
TypeScript
Magic Component Platform (MCP)

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。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

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。

官方
精选
本地
TypeScript
VeyraX

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。

官方
精选
本地
graphlit-mcp-server

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。

官方
精选
TypeScript
Kagi MCP Server

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。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

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。

官方
精选