Daily Stock Analysis

Daily Stock Analysis (DSA) is an open-source, MIT-licensed project by ZhuLinsen that combines multi-source market quotes, news, technicals, and fundamentals with large language models to produce per-stock decision reports and market reviews. It covers A-share, Hong Kong, US and other markets, and can run on a schedule through GitHub Actions, locally, or in Docker. The latest release listed is v3.32.0.

What It Is

DSA is a watchlist research tool. You configure a list of stock symbols, at least one AI model key, and at least one notification channel (or inspect reports through the WebUI or local files). It then gathers quotes, K-line data, technical indicators, news, filings and fundamentals, and has an LLM turn them into a structured report. The project states it is a research system and not investment advice.

Reports and Outputs

• Watchlist decision reports include a score, trend, advice, risk alerts, catalysts, and an action checklist for each symbol.
• Market reviews summarize indices, market breadth, sector strength, and market temperature.
• Reports are stored in SQLite and used for backtests covering direction accuracy, simulated returns, and portfolio review.
• Reports can be pushed to WeChat Work, Feishu, Telegram, Discord, Slack, and email.

Agent Strategy Q&A

The Agent mode supports multi-turn follow-up questions on individual symbols from the Web, CLI, bot, or API entry points. The README lists 15 built-in strategies, including moving averages, Chan theory, and wave theory, and notes experimental custom strategy files and multi-agent orchestration.

Models and Data Sources

Supported model providers include Gemini, OpenAI-compatible endpoints (such as DeepSeek and Qwen), Claude, and Ollama local models. Market data comes from sources such as TickFlow, AkShare, Tushare, Baostock, YFinance, and Longbridge with fallback paths. Free sources run with zero configuration, though the README notes they are subject to upstream rate limits and their stability is not guaranteed. News search can use services such as Anspire, SerpAPI, Tavily, Bocha, Brave, MiniMax, and SearXNG.

Setup Path

There are three runtimes: a serverless GitHub Actions schedule (the recommended option, running weekday evenings by default), a local Python run with main.py and an optional WebUI on port 8000, and Docker via Compose or official Docker Hub and GHCR images. A desktop client packaging guide is also referenced.