QuantGPT
starsforks
How to install QuantGPT?
miasyster/quantgpt
$git clone https://github.com/Miasyster/QuantGPTAgent-Driven Factor Research Engine — Autonomous Mining, Independent Validation on [QuantGPT Cloud](https://quant-gpt.com)
README.md
| 1 | <div align="center"> |
| 2 | |
| 3 | # QuantGPT |
| 4 | |
| 5 | **Agent-Driven Factor Research Engine — Autonomous Mining, Independent Validation on [QuantGPT Cloud](https://quant-gpt.com)** |
| 6 | |
| 7 | LLM Agent 自治因子挖矿 → 批量回测 → 多维评分 → 反过拟合验证 → [QuantGPT Cloud](https://quant-gpt.com) 独立验证 + 样本外跟踪 | 全程零人工干预 |
| 8 | |
| 9 | [](https://github.com/Miasyster/quantgpt/actions/workflows/ci.yml) |
| 10 | [](https://python.org) |
| 11 | [](https://fastapi.tiangolo.com) |
| 12 | [](https://react.dev) |
| 13 | [](LICENSE) |
| 14 | |
| 15 | [Quick Start](docs/QUICKSTART.md) · |
| 16 | [Architecture](docs/ARCHITECTURE.md) · |
| 17 | [API Docs](docs/API_DOC.md) · |
| 18 | [MCP Guide](docs/MCP_GUIDE.md) · |
| 19 | [Factor Mining](docs/FACTOR_MINING.md) · |
| 20 | [Roadmap](ROADMAP.md) · |
| 21 | [Contributing](CONTRIBUTING.md) |
| 22 | |
| 23 | **WeChat / 微信交流群**: Add `quantgpt_ai` on WeChat to join the community |
| 24 | |
| 25 | <a href="https://star-history.com/#Miasyster/QuantGPT&Date"> |
| 26 | <img src="docs/images/star-history.png" width="600" alt="Star History Chart" /> |
| 27 | </a> |
| 28 | |
| 29 | </div> |
| 30 | |
| 31 | --- |
| 32 | |
| 33 | ## What Is QuantGPT |
| 34 | |
| 35 | QuantGPT is an **agent-driven factor research engine** — not a backtest library, not a chatbot wrapper. It gives an LLM Agent (Claude, via MCP) a complete toolkit to autonomously discover, evaluate, iterate, and validate alpha factors, with zero human intervention per research cycle. Factors that pass local validation are automatically uploaded to [**QuantGPT Cloud**](https://quant-gpt.com) for independent verification and out-of-sample tracking — ensuring research results are reproducible and auditable. |
| 36 | |
| 37 | The core architecture: |
| 38 | |
| 39 | ``` |
| 40 | LLM Agent (Claude Code / Claude Desktop) |
| 41 | │ |
| 42 | ├── MCP Tools (15 个) ← Agent 的工具箱 |
| 43 | │ ├── run_backtest ← 全市场分组回测 |
| 44 | │ ├── score_factor ← 0-100 综合评分 |
| 45 | │ ├── diagnose_factor ← 失败模式诊断 |
| 46 | │ ├── run_anti_overfit ← 4 项反过拟合检验 |
| 47 | │ ├── run_rolling_validation ← Walk-forward 验证 |
| 48 | │ ├── validate_expression ← 语法校验 |
| 49 | │ ├── list_operators ← 60+ 算子文档 |
| 50 | │ ├── list_universes ← 股票池和基准 |
| 51 | │ ├── wq_brain_submit ← WQ BRAIN 单因子提交 |
| 52 | │ ├── wq_brain_batch_submit ← 批量参数扫描提交 |
| 53 | │ ├── wq_brain_submit_by_ids ← 按 ID 提交 |
| 54 | │ ├── wq_brain_list_alphas ← 查询已提交 alpha |
| 55 | │ ├── wq_brain_check_alphas ← 检查 alpha 状态 |
| 56 | │ └── wq_brain_finalize_submissions ← 最终提交确认 |
| 57 | │ |
| 58 | ├── QuantGPT Cloud ← 独立验证平台 (quant-gpt.com) |
| 59 | │ ├── A 级因子自动上传 |
| 60 | │ ├── 独立 IC/IR/Turnover 验证 |
| 61 | │ ├── 样本外实时跟踪 |
| 62 | │ ├── 因子去重（Self-Correlation 检测） |
| 63 | │ └── 公开可审计的研究记录 |
| 64 | │ |
| 65 | ├── Evolution Engine ← 因子进化引擎 |
| 66 | │ ├── MutationEngine (8 方向突变) |
| 67 | │ ├── CrossoverEngine (高分因子交叉) |
| 68 | │ ├── MetaEvolutionSelector (自适应策略) |
| 69 | │ └── TrajectoryAnalyzer (轨迹分析) |
| 70 | │ |
| 71 | ├── WQ BRAIN Integration ← WorldQuant 直连（可选） |
| 72 | │ ├── Dollar-neutral 模拟 |
| 73 | │ └── 一键正式提交 |
| 74 | │ |
| 75 | └── Knowledge Base ← 跨会话知识积累 |
| 76 | ├── rules/ (已验证规则) |
| 77 | ├── findings/ (经验发现) |
| 78 | └── failures/ (已证伪路径) |
| 79 | ``` |
| 80 | |
| 81 | ### How It Differs from "AI Backtest Tools" |
| 82 | |
| 83 | 传统工具（包括 ChatGPT + 回测库）的模式是：**人类想因子 → 工具跑回测 → 人类看结果**。Agent 是执行者，人类是决策者。 |
| 84 | |
| 85 | QuantGPT 的模式是：**人类定义目标 → Agent 自治研究 → 人类审阅产出**。Agent 是研究者，人类是审阅者。 |
| 86 | |
| 87 | 这不是接口层的区别（自然语言 vs. 代码），而是**决策权**的区别。Agent 自主决定：探索哪个方向、生成什么表达式、评估哪些指标、何时迭代、何时放弃、何时提交。 |
| 88 | |
| 89 | ### Production Track Record |
| 90 | |
| 91 | | Metric | Value | |
| 92 | |:-------|:------| |
| 93 | | 累计回测任务 | **370+** | |
| 94 | | 单轮迭代（8 候选因子） | **~15 分钟** | |
| 95 | | 表达式算子 | **60+（含非线性、三元、技术指标）** | |
| 96 | | Cloud 独立验证 | **A 级因子自动上传 [quant-gpt.com](https://quant-gpt.com) 验证 + 样本外跟踪** | |
| 97 | | WQ BRAIN 提交（可选） | **3 个因子 IS 全部 PASS，已提交（最佳 Fitness 1.26）** | |
| 98 | |
| 99 | --- |
| 100 | |
| 101 | ## Validated Results |
| 102 | |
| 103 | QuantGPT 产出的因子通过两层验证：本地多维评分 + [QuantGPT Cloud](https://quant-gpt.com) 独立 IC/IR 验证。高质量因子还可选提交至 WQ BRAIN。以下 3 个因子已通过 WQ BRAIN |
People who liked this also viewed
6 picksType
- FinanceToolkitTransparent and Efficient Financial AnalysisMCP ServersJul 2026
- mcp-serverAn MCP server for interacting with the Financial Datasets stock market API.MCP ServersJun 2025
- mcp-serverOfficial Python MCP server for local interactions with the QuantConnect APIMCP ServersMay 2026
- alpaca-mcp-serverAlpaca’s official MCP Server lets you trade stocks, ETFs, crypto, and options, run data analysis, and build strategies in plain English directly from your favorite LLM tools and IDEsMCP ServersJul 2026
- finopsmcpSee where your cloud and AI bills go, and spend less. Runs in your terminal or inside Claude, Cursor, and VS Code. Open source, local-first.MCP ServersJul 2026
- tradingview-mcpTradingView MCP server — real-time market data, technical analysis, screeners & backtesting for Claude, ChatGPT, Cursor & any MCP client. Stocks, crypto, forex & futures across global exchanges. Hosted or self-host.MCP ServersJul 2026
