Plaid MCP Server
An MCP server for integrating with Plaid's API in sandbox mode, enabling linking bank accounts, fetching transactions, balances, and account information.
README
Plaid MCP Server (Python)
A Model Context Protocol (MCP) server for integrating with Plaid's API in sandbox mode. This server provides tools for linking bank accounts, retrieving transactions, balances, and account information.
Features
- Create Link tokens for Plaid Link initialization
- Exchange public tokens for access tokens
- Retrieve account information
- Fetch transactions within date ranges
- Get real-time balance information
Prerequisites
- Python 3.10 or higher
- pip or uv package manager
- Plaid account with sandbox credentials
Setup
1. Get Plaid Credentials
- Sign up at Plaid Dashboard
- Navigate to Developers > Keys
- Copy your
client_idandsandboxsecret
2. Install Dependencies
Using pip:
pip install -e .
Or using uv (faster):
uv pip install -e .
3. Configure Environment
cp .env.example .env
Edit .env and add your credentials:
PLAID_CLIENT_ID=your_client_id_here
PLAID_SECRET=your_sandbox_secret_here
Usage
Running the Server
python -m plaid_mcp_server.server
Or if installed:
plaid-mcp-server
Getting a Sandbox Access Token
For testing, run the sandbox setup script:
python sandbox_setup.py
This will create a test Plaid item and give you an access token to use with the MCP tools.
Available Tools
1. create_link_token
Create a Link token for initializing Plaid Link.
Parameters:
user_id(required): Unique ID for the userclient_name(required): Name of your applicationproducts(optional): Array of products (default: ["transactions"])country_codes(optional): Array of country codes (default: ["US"])language(optional): Language code (default: "en")
2. exchange_public_token
Exchange a public token for an access token.
Parameters:
public_token(required): Public token from Plaid Link
3. get_accounts
Retrieve account information.
Parameters:
access_token(required): Access token for the linked item
4. get_transactions
Fetch transactions within a date range.
Parameters:
access_token(required): Access token for the linked itemstart_date(required): Start date (YYYY-MM-DD)end_date(required): End date (YYYY-MM-DD)
5. get_balance
Get real-time balance information.
Parameters:
access_token(required): Access token for the linked item
Testing in Sandbox
The server is configured to use Plaid's sandbox environment. Use these test credentials when testing with Plaid Link:
- Username:
user_good - Password:
pass_good
See Plaid's Sandbox Guide for more test credentials and scenarios.
Development
Project Structure
plaid-mcp-python/
├── src/
│ └── plaid_mcp_server/
│ ├── __init__.py
│ └── server.py # Main MCP server implementation
├── sandbox_setup.py # Script to create test access tokens
├── requirements.txt # Python dependencies
├── pyproject.toml # Project metadata and build config
├── .env.example # Environment variables template
└── README.md
Code Style
This implementation follows Python best practices:
- Type hints for better IDE support and type checking
- Async/await for concurrent operations
- Proper error handling with specific exceptions
- PEP 8 style guide compliance
- Docstrings for all modules and functions
Connecting to Claude Desktop
To use this MCP server with Claude Desktop, add the following to your Claude Desktop config:
MacOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"plaid": {
"command": "python",
"args": ["-m", "plaid_mcp_server.server"],
"cwd": "/absolute/path/to/plaid-mcp-python",
"env": {
"PLAID_CLIENT_ID": "your_client_id_here",
"PLAID_SECRET": "your_sandbox_secret_here"
}
}
}
}
Or if using uv:
{
"mcpServers": {
"plaid": {
"command": "uv",
"args": ["run", "plaid-mcp-server"],
"cwd": "/absolute/path/to/plaid-mcp-python",
"env": {
"PLAID_CLIENT_ID": "your_client_id_here",
"PLAID_SECRET": "your_sandbox_secret_here"
}
}
}
}
After updating the config, restart Claude Desktop.
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。
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。
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。
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。
mcp-server-qdrant
这个仓库展示了如何为向量搜索引擎 Qdrant 创建一个 MCP (Managed Control Plane) 服务器的示例。