43 professional trading methodologies. One AI that picks the right one for your question.
Traders need many different analysis methods — CANSLIM for growth screening, VCP for breakout patterns, Druckenmiller for strategy synthesis, O'Neil for market timing. But remembering 43 methodologies and knowing which one to use when is hard.
Trading Copilot solves this. Describe your trading idea in plain language (English or Chinese). The AI automatically picks the right methodology and gives you structured, actionable analysis — with entry prices, stop losses, and position sizing.
- No setup required — open the web app and start asking
- Free AI (MiniMax M1) — no API key needed to try
- Live market data (Yahoo Finance) — real prices, not hallucinated numbers
- Works in English and Chinese
- All 43 skills included in this repo — fully self-contained
┌──────────────────────────────────────────────────┐
│ Trading Copilot (Web) │
│ 用户输入交易想法 → AI 自动分析 │
├─────────────────────┬────────────────────────────┤
│ Landing Page │ Chat Mode │
│ 居中输入框 │ 流式 Markdown 渲染 │
│ 6 个快捷按钮 │ 对话历史持久化 │
│ 用户 Profile │ Copy / 时间戳 │
└─────────────────────┴────────────────────────────┘
│ │
▼ ▼
┌──────────────────┐ ┌──────────────────────────┐
│ Vercel Serverless│ │ System Prompt (SKILL.md) │
│ /api/chat.js │ │ 43 trading methodologies │
│ /api/config.js │ │ Intent Routing 表 │
└────────┬─────────┘ │ 输出格式规范 │
│ └──────────────────────────┘
▼
┌──────────────────┐
│ MiniMax M1 (免费) │
│ 或 Anthropic BYOK │
└──────────────────┘
打开 Trading Copilot → 输入你的交易想法 → 得到分析。
# 1. 克隆仓库
git clone https://github.com/zinan92/trading-copilot.git
cd trading-copilot
# 2. 安装依赖
pip install fastapi uvicorn httpx
# 3. 启动（MiniMax 免费 AI）
MINIMAX_API_KEY=your_key python3 app/server.py
# 4. 打开 http://localhost:3000# All 43 skills are included — no external repo needed
git clone https://github.com/zinan92/trading-copilot.git
cd trading-copilot && ./install.sh
# Use in Claude Code
/trading-hub| 功能 | 说明 | 状态 |
|---|---|---|
| Trading Copilot Chat | Web AI trading assistant with auto methodology routing | ✅ |
| 43 Trading Methodologies | CANSLIM, VCP, Druckenmiller, Buffett, O'Neil + more | ✅ |
| Live Market Data | Real-time prices via Yahoo Finance (indexes + stocks) | ✅ |
| Streaming Markdown | Tables, headers, bold rendered in real-time | ✅ |
| Thinking UI | Cursor-style progress (Fetching → Analyzing → Generating) | ✅ |
| User Profile | Trading style/market/risk preferences for personalized analysis | ✅ |
| Dual AI Provider | MiniMax (free) + Anthropic Claude (BYOK) | ✅ |
| Chat Persistence | History saved in localStorage, survives refresh | ✅ |
| Demo Showcase | Pre-generated NVDA analysis, zero API key needed | ✅ |
| 43 Skills in Repo | Fully self-contained, no external dependencies | ✅ |
| Vercel Deployment | Free cloud hosting, accessible from any device | ✅ |
| 分类 | 数量 | 核心方法 |
|---|---|---|
| Macro & Regime | 5 | macro-regime-detector, macro-liquidity, us-market-sentiment, us-market-bubble-detector, market-environment-analysis |
| Market Timing | 6 | market-breadth-analyzer, breadth-chart-analyst, market-top-detector, ftd-detector, uptrend-analyzer, market-news-analyst |
| Stock Screening | 8 | canslim-screener, vcp-screener, pead-screener, dividend-growth-pullback-screener, value-dividend-screener, finviz-screener, theme-detector, institutional-flow-tracker |
| Stock Analysis | 5 | us-stock-analysis, us-value-investing, tech-earnings-deepdive, technical-analyst, btc-bottom-model |
| Strategy & Execution | 8 | stanley-druckenmiller-investment, position-sizer, options-strategy-advisor, backtest-expert, risk-management, pair-trade-screener, scenario-analyzer, trade-hypothesis-ideator |
| Earnings | 3 | earnings-calendar, earnings-trade-analyzer, economic-calendar-fetcher |
| Portfolio | 2 | portfolio-manager, sector-analyst |
| A-Share | 6 | ashare-daily-review, ashare-concept-tracker, ashare-signal-scanner, ashare-stock-screener, ashare-watchlist-briefing, trading-hub |
| 工作流 | 步骤 |
|---|---|
| 晨间复盘 | economic-calendar → market-news → breadth → sector → bubble-detector |
| 选股流程 | screener (CANSLIM/VCP/Theme) → stock-analysis → position-sizer |
| 财报季 | earnings-calendar → tech-earnings-deepdive → earnings-trade-analyzer |
| 策略合成 | 8 upstream skills → druckenmiller synthesizer → 0-100 conviction score |
| A 股复盘 | daily-review → concept-tracker → signal-scanner → watchlist-briefing |
| 层级 | 技术 | 用途 |
|---|---|---|
| 前端 | HTML + CSS + JS (单文件) | Chat 终端 UI |
| 渲染 | marked.js + DOMPurify | Markdown 安全渲染 |
| 字体 | DM Sans + DM Mono | 金融科技风格 |
| 后端(本地) | FastAPI + httpx | API 代理 + 流式转发 |
| 后端(云端) | Vercel Serverless (Node.js) | 零运维部署 |
| AI | MiniMax M1 (免费) / Anthropic Claude | 对话生成 |
| 持久化 | localStorage | 用户 Profile + 聊天记录 |
trading-copilot/
├── app/ # Trading Copilot (web app)
│ ├── index.html # Chat terminal UI
│ ├── server.py # FastAPI local server (with data prefetch)
│ └── prompts/SKILL.md # System prompt with all methodologies
├── api/ # Vercel Serverless Functions
│ ├── chat.js # AI proxy (MiniMax/Anthropic) + Yahoo Finance data
│ └── config.js # Config endpoint
├── skills/ # All 43 trading methodology skills
│ ├── canslim-screener/ # O'Neil CANSLIM growth screening
│ ├── vcp-screener/ # Minervini VCP breakout patterns
│ ├── stanley-druckenmiller-investment/ # Strategy synthesis (0-100 conviction)
│ ├── tech-earnings-deepdive/ # 16-module earnings analysis
│ ├── market-top-detector/ # O'Neil distribution day detection
│ ├── ftd-detector/ # Follow-through day confirmation
│ ├── ... (43 skills total) # See full list in "Built-in Methodologies"
│ └── trading-hub/ # Unified entry point for Claude Code
├── docs/screenshots/ # Product screenshots
├── catalog.html # Skill catalog display page
├── vercel.json # Vercel deployment config
├── install.sh # Claude Code skill installer
└── LICENSE # MIT
| 变量 | 说明 | 必填 | 默认值 |
|---|---|---|---|
MINIMAX_API_KEY |
MiniMax API Key (免费 AI) | 是(二选一) | 无 |
ANTHROPIC_API_KEY |
Anthropic API Key (BYOK) | 是(二选一) | 无 |
FMP_API_KEY |
Financial Modeling Prep (Claude Code skills 用) | 否 | 无 |
不设 API key 时，用户可在网页 Settings 中手动输入自己的 key。
本节面向需要将此项目作为工具或依赖集成的 AI Agent。
name: trading-copilot
description: AI trading analysis terminal with 43 professional methodologies and auto-routing
version: 1.0.0
type: web-app + claude-code-skill-collection
web_app:
url: https://trading-skills-catalog-xbb3.vercel.app/
api_base: https://trading-skills-catalog-xbb3.vercel.app/api
endpoints:
- path: /api/chat
method: POST
description: Send trading question, get streaming AI analysis
body:
message: string (user's trading question)
history: array (conversation history)
provider: string (minimax | anthropic)
api_key: string (optional, for BYOK)
user_profile: string (optional, trading preferences)
response: text/event-stream
- path: /api/config
method: GET
description: Check available AI providers
skill_entry_point:
name: trading-hub
trigger: /trading-hub
description: Menu-driven entry point with 5 workflows and 6 category browsers
install_command: ./install.sh
health_check: curl https://trading-skills-catalog-xbb3.vercel.app/api/config
capabilities:
- auto-route user trading intent to best analysis methodology
- analyze stocks using CANSLIM, VCP, Buffett, Druckenmiller frameworks
- synthesize market breadth, regime, and timing into conviction score (0-100)
- calculate position sizes using Kelly criterion and ATR
- screen stocks by natural language conditions
- analyze tech earnings with 16-module institutional framework
categories:
macro_regime: 5 methods
market_breadth_timing: 5 methods
stock_screening: 10 methods
stock_analysis_reports: 9 methods
strategy_execution: 10 methods
ashare: 5 methodsimport httpx
async def ask_trading_copilot(question: str):
"""Ask Trading Copilot a trading question and get streaming analysis."""
async with httpx.AsyncClient() as client:
async with client.stream(
"POST",
"https://trading-skills-catalog-xbb3.vercel.app/api/chat",
json={"message": question, "provider": "minimax"},
timeout=60
) as resp:
async for line in resp.aiter_lines():
if line.startswith("data: "):
data = json.loads(line[6:])
if data["type"] == "text":
print(data["text"], end="", flush=True)
elif data["type"] == "done":
break
# Example
await ask_trading_copilot("帮我分析一下 NVDA 值不值得买入")| 项目 | 说明 | 链接 |
|---|---|---|
| tradermonty/claude-trading-skills | 30+ 交易分析 Skills 源头 | GitHub |
| Day1Global-Skills | 机构级财报分析 Skill | GitHub |
| AI-Trader | 多 AI 模型实盘竞技 | GitHub |
| TradingAgents | 多 Agent 交易框架 | GitHub |
| daily-stock-analysis | A/H/US 自动分析 + 推送 | GitHub |
MIT