https://github.com/schonarth/pivot
AI-powered paper trading app with news sentiment + technical analysis, simulation, alerts with auto-trade, and MCP integration
https://github.com/schonarth/pivot
ai-analysis papertrade papertrader trade trader tradinganalysis
Last synced: about 2 months ago
JSON representation
AI-powered paper trading app with news sentiment + technical analysis, simulation, alerts with auto-trade, and MCP integration
- Host: GitHub
- URL: https://github.com/schonarth/pivot
- Owner: schonarth
- License: mit
- Created: 2026-04-12T20:58:51.000Z (2 months ago)
- Default Branch: main
- Last Pushed: 2026-04-19T23:43:08.000Z (2 months ago)
- Last Synced: 2026-04-20T01:26:16.488Z (2 months ago)
- Topics: ai-analysis, papertrade, papertrader, trade, trader, tradinganalysis
- Language: Python
- Homepage:
- Size: 949 KB
- Stars: 3
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
- Agents: AGENTS.md
Awesome Lists containing this project
README
#  Pivot
**An AI-powered paper trading simulator** for developing and testing investment strategies across global markets without risking real capital. Build strategies, get AI-generated insights with broad market context, track how the story changes over time, compare portfolio or watchlist summaries, and surface new assets worth watching, all risk-free — or let your AI agent do it for you with MCP support!
---
## What You Can Do
- **Develop strategies** with technical indicators and backtesting engine
- **Get AI insights** — Technical analysis powered by your choice of AI providers, with linked source headlines and broader context beyond the ticker
- **Backtest strategies** against historical OHLCV data
- **Paper trade** across 4 global markets: Brazil, US, UK, and Europe (Your market isn't here? Feel free to add! 😄)
- **Set smart alerts** that trigger trades automatically when conditions are met
- **Simulate market prices** to test strategies when markets are closed
- **Manage multiple portfolios** and watchlists for different investment goals and risk profiles, with AI summaries that can cover the whole set
- **Discover new opportunities** with ranked assets that are worth watching before you own them
- **Bring your AI agent** to trade with you with our simple to setup MCP server
Perfect for developing trading strategies, testing alert logic, learning portfolio management, and experimenting with automated trading — all without risking real capital.
---
## Getting Started
### Requirements
You only need **Docker**. That's it.
**Don't have Docker?**
- [Download Docker Desktop](https://www.docker.com/products/docker-desktop)
- Install and run it
### Start the Application
Once Docker is running:
**Linux/Mac:**
```bash
./start.sh
```
**Windows:**
```bash
start.bat
```
The application will be ready at `http://localhost:3000` in about 30 seconds.
### First Time Setup
1. Open `http://localhost:3000` in your browser
2. Register with an email and password
3. Create your first portfolio (choose a market)
4. Deposit starting capital
5. Start trading!
---
## Features
### Strategy Development
- **Build strategies** with technical indicators (moving averages, RSI, MACD, Bollinger Bands, and more)
- **Backtest** strategies against historical OHLCV data
- **Monitor execution** — Track autonomous trades triggered by your strategies
- **Test and iterate** — Refine rules before deploying to live trading
### AI-Powered Market Intelligence
- **Generate technical analysis** — Claude, GPT, and Gemini supported now for market insights, with recent headlines tied back to the original source articles
- **Follow the narrative** — news digest is kept short and readable, so the insight can explain what changed instead of treating each run like a blank slate
- **Compare scope-level context** — portfolio and watchlist summaries help you see whether the whole set is moving together or diverging
- **Spot opportunities earlier** — discovery surfaces a capped shortlist of assets worth watching, with a concise reason for each one
- **Budget tracking** — Monitor AI provider costs and set spending limits (API key required — sorry, no flat rate support 😅)
### Portfolios
- Create separate portfolios for different markets
- Track cash, invested value, and profit/loss
- View performance history and returns
- See AI summaries for portfolios and watchlists when you want a quick read on the whole group
- Use divergence-style comparisons to see how a holding or watch item differs from the surrounding context
- Deposit or withdraw funds anytime
### Trading
- Buy and sell stocks/ETFs with automatic fee calculation
- (no real money transactions, this is for study)
- Track your positions and average costs
- View current prices and unrealized gains/losses
### Price Alerts
- Set alerts when prices go above or below your target
- Get notified when alerts trigger
- Automatically execute trades when conditions are met
- Choose fixed quantity or percentage-based trades
### Market Simulation
Test your strategies when markets are closed:
**Quick price update:**
```bash
./simulate_price.sh AAPL 150.00
```
**Run realistic market movement:**
```bash
./simulate_market.sh
```
**Simulate with directional trends:**
```bash
./simulate_market.sh --trends AAPL MSFT
```
**Simulate specific market:**
```bash
./simulate_market.sh BR
```
The simulator updates prices in real-time—watch your portfolio and strategies react on the dashboard.
> [!IMPORTANT]
> To enable alerts for simulated market, you must turn on the **simulation toggle** on your portfolios.
> [!TIP]
> Use different portfolios to run simulations and for actual market tracking, to prevent distortions.
---
## AI-Powered Automation (MCP Server)
Use AI agents to automate trading, backtest strategies, generate insights, and manage your portfolio.
**Agent capabilities:**
- Manage trades, alerts, and portfolios
- Backtest trading strategies against historical data
- Generate technical analysis and trading insights
- Monitor strategy execution and autonomous trades
- Manage AI provider settings and cost budgets
- Look up assets through trusted sources when a symbol is missing locally
**To set up agent access:**
1. Go to Settings
2. Copy the prompt template with your User UUID
3. Generate a one-time password (OTP) and give it to your agent
4. Your agent will have full access to manage trading, backtesting, discovery, and insights
---
## Supported Markets
| Market | Currency | Trading Fee | Exchange |
|--------|----------|-------------|----------|
| Brazil | BRL | 0.03% | BVMF |
| United States | USD | 0% | XNYS |
| United Kingdom | GBP | 0.1% | XLON |
| Europe | EUR | 0.1% | XPAR |
---
## Troubleshooting
**"Can't connect to http://localhost:3000"**
- Make sure Docker is running
- Run the start script again
- Wait 30 seconds for services to initialize
- If you're accessing from a different machine than you're running the Pivot server, replace `localhost` with the server's IP address.
**"Application won't load"**
- Check that port 3000 isn't in use by another application
- Restart Docker
**"Simulator not updating prices"**
- Make sure the application is running (`./start.sh` or `start.bat`)
- Try a fresh start: stop the application and run the start script again
**Need to stop the application?**
```bash
docker compose down
```
---
## License
MIT License. See LICENSE file for details.
---
**Built with ❤️ for learning and experimentation.**