yfinance-mcp-server2

Exposes yfinance's stock ticker data fetching through the Model Context Protocol, allowing AI agents to query current and historical stock prices by ticker symbol. Implements MCP resource endpoints that wrap yfinance's Ticker.history() and Ticker.info methods, translating HTTP/JSON requests into structured financial data responses with OHLCV (open, high, low, close, volume) fields and metadata.

Supports querying multiple stock tickers in a single MCP call, aggregating results into a unified response structure. Implements parallel or sequential fetching of yfinance Ticker objects, combining OHLCV data, fundamentals, and metadata across symbols, then normalizing into a consistent JSON schema for downstream processing by AI agents.

Exposes yfinance's Ticker.history() method through MCP, allowing agents to fetch historical price data at configurable intervals (daily, weekly, monthly). Translates interval parameters into yfinance's period/interval arguments, returns time-indexed DataFrames converted to JSON arrays with timestamp, open, high, low, close, volume, and adjusted close fields.

Exposes yfinance's Ticker.info dictionary through MCP, extracting key fundamental metrics (PE ratio, market cap, dividend yield, earnings per share, 52-week high/low, etc.). Implements selective field extraction from yfinance's unstructured info dict, normalizing null/missing values and converting numeric strings to proper types for agent consumption.

Implements the Model Context Protocol server specification, registering yfinance capabilities as MCP resources and tools with proper schema definitions. Uses MCP's JSONSchema for input validation, implements request/response serialization, and handles MCP lifecycle (initialization, capability advertisement, error handling). Enables Claude and other MCP clients to discover and invoke yfinance functions with type-safe arguments.

Implements input validation for ticker symbols and date parameters, catching invalid tickers early and returning structured error responses via MCP. Validates ticker format (alphanumeric, length constraints), date range logic (start < end), and handles yfinance exceptions (network errors, invalid symbols) by translating them into MCP error responses with descriptive messages for agent consumption.

Handles common ticker symbol variations and aliases (e.g., 'Apple' → 'AAPL', 'BRK.A' → 'BRK-A'), normalizing user input before querying yfinance. Implements a mapping layer for common aliases and case-insensitive matching, allowing agents to accept natural language ticker references and convert them to valid yfinance symbols.

Provides a runnable MCP server implementation with standard lifecycle hooks (startup, shutdown, error recovery). Implements stdio-based MCP transport for local execution or can be deployed as a subprocess managed by MCP clients (Claude Desktop, custom hosts). Handles graceful shutdown, resource cleanup, and connection state management for reliable agent integration.