FinAgent MCP
Enables LLMs to access banking, investment, and cryptocurrency data through a standardized MCP interface, featuring Plaid integration for account and transaction data and Robinhood for crypto trading with audit trails.
README
FinAgent MCP
MCP (Model Context Protocol) server for financial data integration, featuring Plaid banking/investment data and Robinhood cryptocurrency trading with advanced context engineering and evidence tracking for a financial data platform that provides LLMs with structured access to banking, investment, and cryptocurrency data through a standardized MCP interface. The system combines Node.js MCP tooling with Go-based data ingestion to deliver real-time financial insights with full observability and audit trails.
Architecture (Diagram via ChatGPT)
[Node MCP Server] ──► [Go Ingestion Service] ──► [External APIs]
• Tool registry • Plaid sync • Plaid API
• Context packer • Robinhood client • Robinhood API
• Evidence builder • Webhook handlers
• Rate limiting • Data encryption
│ │
▼ ▼
[PostgreSQL] [Redis Cache]
Data storage Rate limits
Audit trails Idempotency
Components:
- Node MCP Server: TypeScript-based tool execution and context engineering
- Go Ingestion Service: High-performance data synchronization and API integration
- PostgreSQL: Primary data storage with full audit trails
- Redis: Caching, rate limiting, and idempotency locks
Features
Banking & Investment Tools
- Account management and monitoring
- Transaction analysis with advanced filtering
- Spending summaries and categorization
- Investment holdings and performance tracking
- Investment transaction history
Cryptocurrency Trading
- Real-time crypto portfolio monitoring
- Secure order placement with safety guardrails
- Risk assessment and position sizing limits
- Dry-run mode for safe testing
Other
- Context packing with intelligent data ranking
- Evidence tracking for auditable data lineage
- OpenTelemetry distributed tracing
- Production-ready rate limiting and security
Environment Variables
Create a .env file in the project root:
DATABASE_URL=postgres://postgres:postgres@localhost:5432/finagent?sslmode=disable
REDIS_URL=redis://localhost:6379
PLAID_CLIENT_ID=plaid_client_id
PLAID_SECRET=plaid_secret
PLAID_ENVIRONMENT=sandbox
ROBINHOOD_USERNAME=robinhood_username
ROBINHOOD_PASSWORD=robinhood_password
ENCRYPTION_KEY=32_char_encryption_key
GO_SERVICE_URL=http://localhost:8081
MCP_SERVICE_URL=http://localhost:3001
WEB_SERVICE_URL=http://localhost:3000
JAEGER_ENDPOINT=http://localhost:14268/api/traces
NODE_ENV=development
LOG_LEVEL=info
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。