YUHAI0/fin-agent

What it solves

Fin-Agent provides a natural language interface for complex financial analysis, eliminating the need for users to write code or remember complicated commands to access market data, perform technical analysis, or backtest trading strategies.

How it works

It acts as an AI agent that integrates Large Language Models (such as DeepSeek, Kimi, or local models via Ollama) with the Tushare financial data API. The agent translates user natural language queries into data requests, analyzes the results using the LLM, and can execute specific tools for tasks like stock screening, strategy backtesting, and price monitoring.

Who it’s for

It is designed for investors and financial analysts who want to leverage AI to quickly query stock quotes, analyze financial statements, monitor macro-economic indicators, and test trading strategies without manual data processing.

Highlights

• Natural Language Interaction: Query stock prices, financial reports, and macro data (GDP, CPI, M2) using plain Chinese.
• Intelligent Stock Screening: Automatically filters stocks based on natural language descriptions of fundamental or technical criteria.
• Strategy Backtesting: Built-in engine supporting over 20 strategies (e.g., MACD, RSI, Turtle, Bollinger Bands) with output for returns and drawdowns.
• Portfolio & Alert Management: Supports simulated portfolio tracking and automated price alerts sent via email.
• Personalized Profiles: Remembers user investment styles (e.g., conservative vs. aggressive) to provide tailored suggestions.
• Multi-Model Support: Compatible with various commercial LLMs and local deployments via OpenAI-compatible APIs.