ArvinLovegood/go-stock

What it solves

Go-stock is an AI-powered stock analysis tool designed to simplify the process of monitoring and analyzing global markets (A-shares, Hong Kong, and US stocks) and ETFs. It integrates real-time market data with large language models (LLMs) to provide automated summaries, sentiment analysis, and intelligent stock selection, reducing the manual effort required to parse financial news and complex charts.

How it works

The tool is built using Wails and NaiveUI, connecting to various LLM providers (such as OpenAI, DeepSeek, and local models via Ollama/LMStudio) to process financial data. It employs an agentic architecture with three modes—React, PlanExecute, and DeepAgents—to execute complex tasks. It uses a skill system and MCP (Model Context Protocol) extensions to expand its capabilities, and integrates a vector database (chromem-go) for long-term memory and knowledge base management.

Who it’s for

It is intended for stock market investors and researchers who want to AI-enhanced technical and fundamental analysis, automated market monitoring, and a centralized dashboard for global stock data.

Highlights

• Multi-Agent Intelligence: Supports React, PlanExecute, and DeepAgents modes for complex task planning and execution.
• Comprehensive Market Coverage: Real-time data for A-shares, Hong Kong, and US stocks, including K-line analysis, fund flow, and ETF tracking.
• Extensible Tooling: Includes over 150 AI data tools, a skill sharing community (Skill Plaza), and MCP OAuth authorization.
• Advanced Analysis: Features automated sentiment analysis with frequency weighting, wave theory drawing tools, and AI-generated analysis reports.
• Integration: Supports notifications via Feishu and DingTalk, and integrates with a wide array of LLM platforms and local model runners.