Analysis updated 2026-05-18
Automatically scan all A-share stocks each evening for breakout and reversal patterns.
Get daily trading signal alerts pushed to a Feishu group chat.
Run a Turtle Trading breakout or moving-average-volume screen across the whole market.
Backfill and store years of Chinese stock price history locally in SQLite.
| sngyai/sequoia-x | endernewton/tf-faster-rcnn | galaxy-dawn/claude-scholar | |
|---|---|---|---|
| Stars | 3,657 | 3,659 | 3,661 |
| Language | Python | Python | Python |
| Setup difficulty | moderate | hard | moderate |
| Complexity | 3/5 | 5/5 | 2/5 |
| Audience | data | researcher | researcher |
Figures from each repo's GitHub metadata at analysis time.
Initial backfill of roughly 5,200 stocks takes about 12 minutes and needs a Feishu webhook configured.
Sequoia-X is a stock screening system built for the Chinese A-share market. It runs automatically after the market closes each trading day, scans all publicly listed stocks for specific technical patterns, and pushes the results to a group chat via the Feishu (Lark) messaging platform. The goal is to find stocks showing recognizable chart setups that traders often look for before deciding to buy. The system includes six built-in screening strategies. These include a Turtle Trading breakout that looks for stocks hitting 20-day highs with strong volume, a moving-average-plus-volume breakout, a high tight flag pattern, a limit-up pullback confirmation, a limit-down reversal during an uptrend, and a relative price strength breakout inspired by William O'Neil's work. Each strategy applies its own rules to the price and volume history of every stock in the market. For market data, the system pulls daily price history from Baostock, a free Chinese financial data service that requires no registration and has no rate limits. Data is stored locally in a SQLite database on the user's machine. The first time you set it up, a backfill run downloads the full historical price data for roughly 5,200 stocks, which takes about 12 minutes. After that, the daily run uses eight parallel processes to fetch only the new data, completing a full market update in two to three minutes. Setup requires Python 3.10 or newer. You install the dependencies, copy an environment variable template and fill in your Feishu webhook URL, run the initial backfill, and then configure a scheduled task to run the script automatically each weekday evening. The project is MIT licensed and the README is written in both Chinese and English.
A Python system that scans the Chinese stock market daily for technical chart patterns and sends alerts to a Feishu chat.
Mainly Python. The stack also includes Python, SQLite, Baostock.
You can freely use, modify, and redistribute this project under the MIT license.
Setup difficulty is rated moderate, with roughly 1h+ to a first successful run.
Mainly data.
This repo across BitVibe Labs
Verify against the repo before relying on details.