personal-capital-connector-mcp
Enables Claude to query Empower/Personal Capital financial accounts, net worth, transactions, and asset allocation via natural language.
README
Personal Capital MCP Connector
A Model Context Protocol (MCP) server that connects Claude to your Empower / Personal Capital financial data. Ask Claude natural-language questions about your accounts, net worth, spending, and investment holdings — all answered from live data.
What it does
| Tool | Example questions |
|---|---|
list_accounts |
"What's my Chase credit card balance?" / "Show my savings accounts" |
get_net_worth |
"What's my net worth?" / "How much do I owe vs own?" |
get_transactions |
"What did I spend at restaurants last month?" |
get_asset_allocation |
"What's my asset allocation in my 401k?" |
check_auth_status |
"Is my Empower session still valid?" |
Project structure
personal-capital-connector-mcp/
├── pyproject.toml
└── src/personal_capital_connector/
├── __init__.py
├── __main__.py
├── auth.py # browser-based 2FA login + session persistence
├── client.py # API wrapper and data formatters
├── server.py # FastMCP server exposing 5 tools to Claude
└── cli.py # CLI entry point (auth / status / serve)
Prerequisites
- uv — fast Python package manager
- An Empower / Personal Capital account
Setup
Step 1 — Clone and authenticate (interactive browser + 2FA):
git clone https://github.com/meetvanani/personal-capital-connector-mcp.git
cd personal-capital-connector-mcp
uv run personal-capital-connector auth
A browser window opens, auto-fills your credentials, and waits for you to complete 2FA. The session is saved to ~/.config/personal-capital-connector/session.json (chmod 600). Re-run this any time your session expires.
Step 2 — Add to Claude Desktop's MCP config:
Open ~/Library/Application Support/Claude/claude_desktop_config.json and add:
{
"mcpServers": {
"personal-capital": {
"command": "uv",
"args": [
"run",
"--directory",
"/absolute/path/to/personal-capital-connector-mcp",
"personal-capital-connector"
]
}
}
}
Replace /absolute/path/to/personal-capital-connector-mcp with the actual path where you cloned the repo.
Step 3 — Restart Claude Desktop and start asking questions.
Other CLI commands
# Check if the saved session is still valid
uv run personal-capital-connector status
# Explicitly start the MCP server (same as default / no subcommand)
uv run personal-capital-connector serve
How authentication works
Login uses Playwright to open a real Chromium browser:
- Auto-fills your email and password
- Waits up to 3 minutes for you to complete 2FA (SMS, email, or authenticator app)
- Detects successful authentication by intercepting the first valid API response from Empower
- Extracts and saves the session cookies + CSRF token to disk
This approach handles Empower's frequent login-flow changes without brittle scraping.
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。