This is an early release preview. You may encounter bugs.
Daily Stock Analysis logo
Unclaimed

Tool research connectors

Daily Stock Analysis

Daily LLM analysis of multi-market stocks producing decision dashboards pushed to chat and email channels

A+ 91/100 GitHub score ? This grade is derived from GitHub signals, not user votes. Open for the full breakdown.
Discussion

01 / Summary

About Daily Stock Analysis.

Daily Stock Analysis (DSA) analyses a watchlist of stocks with a large language model and pushes a daily decision dashboard to chat and email channels. It covers A-shares, Hong Kong, US, Japanese, Korean, and Taiwanese stocks plus ETFs, and the interface and reports are primarily in Chinese.

Each run gathers quotes, candlestick data, technical indicators, news, announcements, and fundamentals from several data sources, then has the model write a report per stock: a core conclusion, a score, the trend, buy and sell levels, risk alerts, catalysts, and an action checklist. A market review summarizes the main indices, breadth, and sector moves. Free market-data sources (AkShare, Baostock, YFinance) work without configuration, and token-based sources such as TickFlow, Tushare, and Longbridge add stability for scheduled or batch runs.

DSA runs on a schedule through GitHub Actions without a server (by default on weekday evenings Beijing time, skipping market holidays), or locally, in Docker, or as a FastAPI service. A web and desktop workspace handles manual analysis, task progress, report history, backtests, holdings, and configuration.

An agent mode answers follow-up questions about a stock over several turns, applying one of 15 built-in strategies (moving-average crossovers, Chan theory, Elliott waves, trend, hot topics, event-driven, growth, and others) and calling live quotes, indicators, news, and risk data. It is available in the web chat, bots, and the API, with custom strategy files and experimental multi-agent orchestration. The project states that it is for study and research and is not investment advice.

  • AI decision reports: conclusion, score, trend, buy and sell levels, risk alerts, catalysts, and an action checklist per stock
  • Multi-market data: A-shares, Hong Kong, US, Japanese, Korean, and Taiwanese stocks and ETFs, with quotes, indicators, news, announcements, and fundamentals
  • Market data sources: TickFlow, AkShare, Tushare, Pytdx, Baostock, YFinance, and Longbridge
  • News search: Anspire, SerpAPI, Tavily, Bocha, Brave, MiniMax, and SearXNG, plus optional Reddit, X, and Polymarket sentiment for US stocks
  • Model providers: Gemini, Claude, DeepSeek, Qwen, OpenAI-compatible APIs, Ollama local models, and aggregators such as Anspire and AIHubMix
  • Agent strategy chat: multi-turn questions with 15 built-in strategies, live data calls, session export, and background execution
  • Notifications: WeCom, Feishu, Telegram, Discord, Slack, and email
  • Scheduling and deployment: GitHub Actions without a server, local scheduled tasks, Docker, or a FastAPI service
  • Web and desktop workspace: manual analysis, report history, full Markdown reports, backtests, holdings, and configuration management
  • Smart import: watchlists from images, CSV or Excel, or the clipboard, with code, name, pinyin, and alias completion

02 / Discussion CREDIBILITY-GATED

Discussion

Reading is open to everyone. Posting and voting need a verified identity or a GitHub grade of B or higher.

  • No discussions yet.

04 / Build

Build with Daily Stock Analysis.

Browse the catalog for frameworks, tools, and harnesses, each scored on real GitHub credibility.

Get Daily Stock Analysis →

Browse the catalog