GitHub Actions as your free quant analyst

What it does This Python system aggregates market data, news, and fundamentals for A-shares, Hong Kong, and US stocks, then prompts an LLM to generate a daily decision dashboard with ratings, risk alerts, and catalysts. It is designed to run on GitHub Actions at zero server cost, automatically pushing reports to enterprise WeChat, Feishu, Telegram, Discord, Slack, or email every weekday evening. A local FastAPI web UI and desktop workbench support manual analysis, backtesting, position tracking, and an agent chat for strategy questions.

The interesting bit Rather than treating GitHub as mere storage, the project uses GitHub Actions as the production cron scheduler—fork the repo, fill in API secrets, and the repository wakes itself up daily at 18:00 Beijing time to analyze your watchlist. It also packs fifteen built-in strategies—from moving average crosses to Elliott wave and Chan theory—into an agent interface that can query live quotes and news in multi-turn conversations.

Key highlights

• Pulls quotes, K-lines, technical indicators, capital flow, and news across multiple markets via AkShare, Tushare, YFinance, TickFlow, and other sources
• Supports a broad range of LLM backends including OpenAI-compatible APIs, DeepSeek, Gemini, Claude, Tongyi Qianwen, and local Ollama instances
• Pushes formatted decision dashboards and market reviews to six notification channels without requiring dedicated infrastructure
• Includes a web-based workbench with light/dark themes, historical reports, CSV/image import, and an agent chat for strategy Q&A
• Offers backtesting and position management alongside automated daily scheduling

Caveats

• Report quality depends heavily on configured news search APIs; the README explicitly warns that sentiment, announcements, and catalyst analysis degrade without at least one search source
• Social sentiment data is limited to US stocks via Reddit, X, and Polymarket, so A-share and HK sentiment relies on news inference rather than direct social scraping
• Getting started requires API keys for at least one LLM provider, one notification channel, and a manually defined stock list, with optional but recommended news search keys

Verdict Worth a look if you want a low-maintenance, multi-market quant briefing in your work chat every evening. Skip it if you need an automated trading bot; this analyzes and notifies, but does not place orders.