Summary

Fin-Agent is an intelligent financial analysis assistant designed for individual investors and financial professionals. It leverages large language models (LLMs) and Tushare financial data to provide natural language-driven insights, stock screening, strategy backtesting, and investment recommendations, aiming to simplify complex financial analysis. A desktop GUI version is also available for enhanced usability.

How It Works

Fin-Agent integrates advanced LLMs with Tushare Pro financial data. Users interact via natural language, enabling data querying, analysis, and insight generation. The system supports a wide array of domestic and local LLMs for deployment flexibility. Core functionalities like strategy backtesting and intelligent stock screening are driven by dedicated engines translating natural language requests into executable financial operations.

Quick Start & Requirements

Installation is straightforward via pip: pip install fin-agent. To run, execute fin-agent in the terminal. Prerequisites:

• A Tushare Token is required for accessing financial data.
• LLM configuration: API keys for commercial models or Base URLs for local models (e.g., Ollama).
• The setup process includes an interactive guide for API key configuration.
• A desktop version is available for a more user-friendly graphical interface.
• For continuous price monitoring, the --worker mode is recommended.
• Website: fin-agent.chat

Highlighted Details

• Natural Language Interface: Enables users to query stock行情, financial data, market indicators, and perform complex analyses using conversational Chinese.
• Advanced Analysis Tools: Includes intelligent stock screening (especially for "long-tail" opportunities), a robust strategy backtesting engine supporting over 20 built-in strategies (e.g., MA cross, MACD, Donchian, Turtle), portfolio management, and real-time price alerts via email.
• Comprehensive Data & Model Support: Integrates deeply with Tushare Pro for a wide range of financial and macroeconomic data. Supports numerous domestic LLMs (DeepSeek, Kimi, Zhipu, Qwen, Yi) and local models via Ollama or LM Studio.
• Personalization: Features user profiling to remember investment styles (conservative/aggressive) and preferences, offering tailored advice.

Maintenance & Community

The README does not explicitly detail maintenance schedules, notable contributors, sponsorships, or community channels like Discord/Slack. The project is actively updated, as indicated by the "Last Commit" badge.

Licensing & Compatibility

The project is licensed under the MIT License. This permissive license generally allows for commercial use, modification, and distribution, making it compatible with closed-source applications.

Limitations & Caveats

The core functionality relies on external API keys for Tushare and LLMs, which may incur costs. The effectiveness of financial advice and analysis is dependent on the quality of the underlying LLM and data. The "worker" mode for continuous monitoring requires the application to remain running.