jundizhou/easy-stock

What it solves

It transforms the fragmented and unstructured nature of A-share market research into a structured, AI-driven workflow. Instead of manually tracking multiple news sources, social media posts from market influencers (Big Vs), and raw price data, it provides a unified workbench that identifies market structures, themes, and sentiment trends with evidence-based AI analysis.

How it works

The system integrates real-time market data from multiple sources (e.g., East Money, Sina) and uses an AI Agent platform called Hermes to process this information. It employs a "Local First" architecture to store research, articles, and session history locally. Key mechanisms include:

• Automated Collection: Agents use a built-in browser to automatically discover and archive articles from platforms like Xueqiu and Taoguba.
• Domain Modeling: It converts raw data into computable models for trends, themes, and sentiment cycles.
• Consensus Extraction: AI analyzes multiple influencer viewpoints to identify common directions, disagreements, and conditions for the next trading day.
• Customized Analysis: It combines market data with user-provided trading philosophies to generate personalized stock reports.

Who it’s for

Investors and researchers focusing on the A-share market who want to move beyond basic price charts to a more systematic, AI-assisted approach to sentiment and theme analysis.

Highlights

• Automated Influencer Review: Automatically collects and summarizes viewpoints from various market authors to find a "daily consensus."
• Short-term Sentiment Analysis: Tracks limit-up ladders, promotion rates, and sentiment cycles in a single view.
• Theme Radar: Identifies market mainlines by aggregating theme rankings, capital flow, and industry strength.
• Evidence-based AI: Ensures AI conclusions are traceable back to original sources, timestamps, and raw data.
• Multi-Model Support: Compatible with OpenAI, DeepSeek, Claude (Anthropic), and other OpenAI-protocol services via the Hermes runtime.