Open-source project
charliedream1/ai_quant_trade avatar
charliedream1/ai_quant_trade

ai_quant_trade is a collection of separate quant sub-projects, each with its own requirements file and its own claims to check

An AI stock trading platform spanning learning, simulation and live trading in one place. Includes stock market knowledge, strategy examples, LLMs, factor mining, traditional strategies, machine learning, deep learning, reinforcement learning, graph networks, high-frequency trading, C++ deployment, and JoinQuant sample code, for easy learning, simulation and live trading.

6,752 stars1,252 forksJupyter NotebookApache-2.0

At a glance

What is it?
charliedream1/ai_quant_trade bundles trading notes, alpha mining, reinforcement learning and LLM examples, a monitoring tool and a curated resource list under one Apache-2.0 repository. It is closer to a teaching library of independent experiments than a single trading platform, and the headline return figures come from individual tutorials.
Who is it for?
ai_quant_trade suits learners who want worked examples across many quant techniques in one place, especially readers of Chinese, since much of the folder structure and some course material is in Chinese. It does not suit anyone looking for one installable trading system or for verified performance: there is no root dependency file, only one tagged release (v0.0.1), and figures such as the 53% annualized return belong to a single tutorial.
Can I use it commercially?
Yes. Apache-2.0 is a permissive licence: you can use, modify and sell software built on it, as long as you keep its copyright and licence notices.
Is it still maintained?
Yes. The repository last received commits 6 days ago.
What is it written in?
Mainly Jupyter Notebook, according to GitHub's language statistics.

Answers come from the project's GitHub data, last synced on November 15, 2026, and from our analysis. They are not legal advice.

From daily snapshots

Star history

  • charliedream1/ai_quant_trade6,752 ★ · +186
Star history
Repository2026-09-232026-11-15
charliedream1/ai_quant_trade6.6k6.8k

186 Stars in the last 53 days, measured from the daily GitHub snapshot. Compare with other projects

Editorial analysis

There is no root install: every egs_ sub-project carries its own requirements.txt

The repository has no global dependency file at its root. Each egs_ directory is a separate sub-project with its own requirements.txt, and the intended workflow is to pick one, change into its directory, and install from there:

sh
pip install -r requirements.txt

The practical consequence is that you never install ai_quant_trade as a whole. You install one example at a time. Two sub-projects can pin different versions of the same library, so running a reinforcement learning example and an LLM example side by side calls for separate virtual environments. It also means there is no single import path or shared package that ties the examples together; code reuse between sub-projects is something you would have to arrange yourself.

The one-stop platform is a set of independent folders added over four years

The project presents itself as covering learning, simulation and live trading in one place, and lists target audiences from institutional investors to retail traders with no programming background. The top level shows how that is organised: ai_notes for trading knowledge in Markdown and notebooks, egs_alpha for alpha mining, egs_trade for strategies and live trading, egs_llm for language model work, egs_fin_nlp for financial text analysis, egs_aide for assistant tools, egs_courses, egs_data, egs_online_platform, egs_skill, plus snippets, tools, runtime and backup directories.

The dated feature log shows these pieces arriving separately: a basic backtest framework with a double moving average strategy in August 2022, Wind live trading simulation in November 2022, a local deep RL strategy in January 2023, StructBERT sentiment analysis in April 2023, LLM market analysis in January 2025, and a V2 stock monitoring tool in July 2026. Each example stands on its own, so older ones may not match the code style or library versions of newer ones.

Headline numbers like 53% annualized return belong to single tutorials, not the repository

Several feature entries carry performance figures in their titles. The March 2023 entry is titled RL Multi-Stock Trading: 53% Annualized Return and points to egs_trade/rl/a002_finRL_tutorial/a01_Stock_NeurIPS2018. The February 2023 entry claims machine learning auto-mining of 5,000 alphas, under egs_alpha/auto_alpha/tsfresh. The August 2025 course on a reasoning stock price forecasting LLM claims a 20% accuracy boost.

The front page gives these as headlines without the test period, market, costs or benchmark behind them. A return figure from one backtest in one tutorial says nothing about how any other strategy in the repository performs, and nothing about live results. Anyone tempted to trade on these examples should rerun the specific notebook, read how its backtest is set up, and treat the headline number as a claim to verify rather than an expected outcome.

Some guides and models sit on the paid Knowledge Planet, not in the repo

Parts of the material are split between GitHub and the author's Knowledge Planet community (AI智投星球) and WeChat account. The May 2025 Unsloth reasoning stock forecasting LLM entry states that the code is in the repository but the detailed guide and the model are on the Planet. The January 2025 LLM financial market analysis entry offers its video tutorial on the Planet or the WeChat account.

For a reader, this means cloning the repository does not always give you everything needed to reproduce a result. With the Unsloth example you get training code but not the trained model or the full walkthrough. Before investing time in an example, check its feature entry or folder for a reference to the Planet. The repository also has a Support Me section and a sponsorship-style link block at the top, which sets expectations about where the deeper material is offered.

Live trading examples depend on outside terminals and platforms

The live trading side does not connect to a broker by itself. The November 2022 entry is a Wind local live trading simulation under egs_trade/real_bid_simulate/wind, which requires access to Wind. The project description also mentions JoinQuant sample code, and egs_online_platform holds material for online research platforms. The assistant tools in egs_aide include an Excel-based stock monitor from February 2023 and its July 2026 successor, the Stealth Stock Trading tool V2, which adds a modular monitoring system, alert monitoring, K-line charts and multi-source fallback.

The core highlights mention coverage of stocks, funds and cryptocurrencies, and deployment through Python, C++, CPU and GPU. The front page does not say which example covers which market or which deployment path, so you need to read each sub-project to find out whether it fits your broker, data source or account type.

One tagged release in a repository that is still changing

The only GitHub release is v0.0.1, published on 2025-02-15. The last push to the master branch was on 2026-10-27, and the newest feature entry is dated 2026-07-25. Change history is kept in a changelog directory and an update_log.md file rather than in tagged releases.

The consequence is that you cannot depend on a version number to know what you are running. An example you clone today may differ from the same folder next month, and the v0.0.1 tag predates the 2025 and 2026 additions. If you build anything on top of a sub-project, record the commit hash you used, because there is no later release to pin to.

Lint, secret scanning and docs configs are present for both Python and C++

The top level holds configuration for several quality tools: .flake8 and .pylintrc for Python, .clang-format and CPPLINT.cfg for C++, a .pre-commit-config.yaml, and a .gitleaks.toml for scanning commits for leaked secrets. There is a unit_test directory, and documentation is set up with mkdocs.yml and .readthedocs.yml alongside the docs folder. CONTRIBUTING.md and CODE_OF_CONDUCT.md are present, as are English, Chinese and Japanese READMEs.

The C++ configs fit the project's stated support for C++ deployment. What these files show is that tooling exists; they do not show how much of the repository it covers. A notebook-heavy repository, with Jupyter Notebook as its primary language, can pass Python linters on scripts while notebooks go unchecked. If you are evaluating code quality for reuse, run the tests in unit_test against the sub-project you care about. The license is Apache-2.0.

Editorial conclusion

ai_quant_trade suits learners who want worked examples across many quant techniques in one place, especially readers of Chinese, since much of the folder structure and some course material is in Chinese. It does not suit anyone looking for one installable trading system or for verified performance: there is no root dependency file, only one tagged release (v0.0.1), and figures such as the 53% annualized return belong to a single tutorial. Before running any example, read that sub-project's own requirements.txt and check whether its full guide sits on the paid Knowledge Planet.

Frequently asked questions

How do I install ai_quant_trade?

ai_quant_trade has no root dependency file. Choose an egs_ sub-project, change into its directory, and run pip install -r requirements.txt there.

Does ai_quant_trade guarantee the 53% annualized return?

No. The 53% annualized return is the title of one reinforcement learning multi-stock tutorial from March 2023, based on the FinRL NeurIPS2018 stock example, not a figure for the repository as a whole.

Is all ai_quant_trade content free on GitHub?

Not all of it. The Unsloth stock forecasting LLM has its code in the repository, but its detailed guide and model are on the author's Knowledge Planet, and some video tutorials are on the Planet or WeChat account.

Official sources

  1. charliedream1/ai_quant_trade on GitHub
  2. Issues
  3. License: Apache-2.0
  4. README
  5. Releases
Add this badge to your README

If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.

Add this badge to your README

markdown
[![Hysen Labs](https://hysenlabs.com/badge/charliedream1-ai-quant-trade.svg)](https://hysenlabs.com/projects/charliedream1-ai-quant-trade)

Star history chart for your README

markdown
[![Star history](https://hysenlabs.com/chart/charliedream1-ai-quant-trade.svg)](https://hysenlabs.com/projects/charliedream1-ai-quant-trade)

Was this useful?