AI column

Data@donghyeok99Staff

daily_stock_analysis: a watchlist report every evening

One user put it on a NAS and now gets a stock analysis pushed to chat every evening

At 18:00 four watchlist tickers come in, a candlestick chart draws itself, and a dashboard message arrives reading Hold 65, Hold 48, Sell 35.
At 18:00 four watchlist tickers come in, a candlestick chart draws itself, and a dashboard message arrives reading Hold 65, Hold 48, Sell 35.
Who
Made by ZhuLinsen for individual investors who can't review every stock on their watchlist each day.
When
Public since January 2026; the latest release is v3.35.0 from November 16, 2026.
Where
Runs on GitHub Actions, your own PC (Python 3.10+) or Docker. Results arrive on Telegram, Slack, Discord, email and more.

What was hard

Even a short watchlist means opening prices, news, filings and money flow one stock at a time after every close. On a busy day you skip it, and after a few skipped days it gets hard to remember why you bought.

What it does about it

You list stock codes once. At a set time it collects quotes and news for each one, hands them to an LLM and sends back a single message with a verdict, a score, price levels, risks and catalysts. The key idea is that no server is needed: a scheduled GitHub Actions run does the work. The same code on your PC or in Docker adds a web workspace where you can start an analysis by hand and reread past reports.

The README's GIF of the web workspace
The README's GIF of the web workspace

What people made with it

sspai.com羊刀仙2026-03-04

A write-up of running it on a QNAP NAS with Docker Compose, scheduled for 19:00 every day. On a free model tier, analyzing and pushing three stocks took close to half an hour.

不过我设置的19:00固定执行,但是碍于硅基流动免费额度的响应速度有限,三只股票分析完并推送需要近乎半小时。

How it runs inside

  1. 1

    GitHub Actions → main.py

    cron '0 10 * * 1-5' → python main.pyMonday to Friday at 10:00 UTC (18:00 Beijing) the workflow sets up Python 3.11 and runs main.py..github/workflows/00-daily-analysis.yml
  2. 2

    main.py → AkShare · YFinance quotes

    get_daily_data · get_realtime_quoteFor each stock it fetches daily bars and a live quote, saves them to its database and computes the trend.src/core/pipeline.py
  3. 3

    main.py → News search API

    search_comprehensive_intelWhen a search service is configured, it gathers the stock's news plus risk and catalyst items.src/search_service.py
  4. 4

    main.py → LLM

    analyzer.analyzeQuotes, trend and news go to the LLM as one context; a verdict and a score come back.src/analyzer.py
  5. 5

    main.py → Dashboard message

    generate_dashboard_report → notifier.sendAll stocks are merged into one dashboard, sent to the channels you set and saved as a report file.src/notification.py

What goes in, what comes out

STOCK_LIST=600519,hk00700,AAPL,7203.T,005930.KS,2330.TWA dashboard message with a verdict, score and trend per stock
python main.py --market-reviewA market review: main indices, advancers and decliners, leading and lagging sectors

Is it for you

Good for

  • Individual investors who want the same evening check on their watchlist every day
  • Anyone who wants to try a scheduled job on GitHub Actions without a server
  • People who want their own analysis screen on a home NAS or in Docker

Not for

  • The README says it is for learning and research and is not investment advice; losses are on the user.
  • You bring one LLM API key and one notification channel yourself.
  • The README notes the default free data sources can be unstable under rate limits and upstream changes.
  • The default README is in Chinese. One user reports close to half an hour for three stocks on a free model tier.

Reviewed by Devalley’s editor before publishing. How the column is made

Sign in to post, comment and like.Sign in

Report

Reason

Read the community rules