Visit the EvoTraders website
EvoTraders is an open-source financial trading agent framework that builds a trading system capable of continuous learning and evolution in real markets through multi-agent collaboration and memory systems.
Before trading with real funds, please conduct thorough testing and risk assessment. Past performance does not guarantee future returns. Investment involves risks — please make decisions with caution.
Core Features
Multi-Agent Collaborative Trading
A team of 6 members — including 4 specialized analyst roles (fundamentals, technical, sentiment, valuation), a portfolio manager, and a risk management agent — collaborating to make decisions like a real trading team.
Continuous Learning and Evolution
Based on the ReMe memory framework, agents reflect and summarize after each trade, preserving experience across rounds and forming unique investment methodologies. Through this design, agents gradually develop their own trading styles and decision preferences rather than producing one-time random inferences.
Real-Time Market Trading
Supports real-time market data integration with both backtesting mode and live trading mode, allowing agents to learn and make decisions amid real market fluctuations.
Visualized Trading Information
Observe agents’ analysis processes, communication records, and decision evolution in real-time, with complete tracking of return curves and analyst performance.
Quick Start
Installation
# Clone repository
git clone https://github.com/agentscope-ai/agentscope-samples
cd agentscope-samples/EvoTraders
# Install dependencies (recommended: uv)
uv pip install -e .
# or: pip install -e .
# Configure environment variables
cp env.template .env
.env file and fill in your API keys. The following variables are required:
# Finance data API
# At minimum, FINANCIAL_DATASETS_API_KEY is required (FIN_DATA_SOURCE=financial_datasets)
# Recommended: add FINNHUB_API_KEY (FIN_DATA_SOURCE=finnhub)
# Live mode requires FINNHUB_API_KEY
FIN_DATA_SOURCE= # finnhub or financial_datasets
FINANCIAL_DATASETS_API_KEY= # Required
FINNHUB_API_KEY= # Optional
# LLM API for agents
OPENAI_API_KEY=
OPENAI_BASE_URL=
MODEL_NAME=qwen3-max-preview
# LLM & embedding API for memory
MEMORY_API_KEY=
Running
Backtest Mode:
evotraders backtest --start 2025-11-01 --end 2025-12-01
evotraders backtest --start 2025-11-01 --end 2025-12-01 --enable-memory # with memory
wget "https://agentscope-open.oss-cn-beijing.aliyuncs.com/ret_data.zip"
unzip ret_data.zip -d backend/data
evotraders live # Run immediately (default)
evotraders live --enable-memory # Run with memory enabled
evotraders live --mock # Mock mode for testing
evotraders live -t 22:30 # Run daily at 22:30 local time (auto-converts to NYSE timezone)
evotraders --help # Global CLI help
evotraders backtest --help # Backtest mode parameters
evotraders live --help # Live/mock mode parameters
# Ensure npm is installed; then run:
evotraders frontend # Connects to port 8765 by default
# Modify ./frontend/env.local to change the port
http://localhost:5173/ to open the trading room. Select a date and click Run or Replay to observe the decision-making process.
System Architecture
Agent Design
Analyst Team
| Role | Focus |
|---|
| Fundamentals Analyst | Financial health, profitability, growth quality |
| Technical Analyst | Price trends, technical indicators, momentum analysis |
| Sentiment Analyst | Market sentiment, news sentiment, insider trading |
| Valuation Analyst | DCF, residual income, EV/EBITDA |
- Portfolio Manager — Integrates analysis signals from all analysts, executes communication strategies, and combines analyst/team historical performance with recent investment memories and long-term experience to make final decisions.
- Risk Management — Monitors real-time price and volatility, enforces position limits, and issues multi-layer risk warnings.
Decision Process
Real-time Market Data → Independent Analysis → Intelligent Communication (1v1 / 1vN / NvN)
→ Decision Execution → Performance Evaluation → Learning & Evolution (Memory Update)
Analysis
Each agent independently analyzes the market using their respective tools and historical experience.
Communication
Agents exchange views through private chats, broadcasts, and group meetings.
Decision
The portfolio manager makes a comprehensive judgment and issues final trades.
Evaluation
- Performance Charts — Track portfolio return curves against benchmark strategies (equal-weighted, market-cap weighted, momentum).
- Analyst Rankings — View win rates across bull and bear markets to identify top contributors.
- Statistics — Detailed position and trading history for in-depth analysis.
Review
Agents reflect on decisions and summarize experiences based on actual returns, storing insights in the ReMe memory framework for continuous improvement. Module Support
| Module | Project |
|---|
| Agent Framework | AgentScope |
| Memory System | ReMe |
| LLM Support | OpenAI, DeepSeek, Qwen, Moonshot, Zhipu AI, and more |
Custom Configuration
Custom Analyst Roles
Step 1. Register the role in ./backend/agents/prompts/analyst/personas.yaml:
comprehensive_analyst:
name: "Comprehensive Analyst"
focus:
- ...
preferred_tools: # Select flexibly based on situation
description: |
As a comprehensive analyst ...
./backend/config/constants.py:
ANALYST_TYPES = {
"comprehensive_analyst": {
"display_name": "Comprehensive Analyst",
"agent_id": "comprehensive_analyst",
"description": "Uses LLM to intelligently select analysis tools, performs comprehensive analysis",
"order": 15
}
}
./frontend/src/config/constants.js:
export const AGENTS = [
{
id: "comprehensive_analyst",
name: "Comprehensive Analyst",
role: "Comprehensive Analyst",
avatar: `${ASSET_BASE_URL}/...`,
colors: { bg: '#F9FDFF', text: '#1565C0', accent: '#1565C0' }
}
]
Custom Models
Configure the model used by each agent in your .env file:
AGENT_SENTIMENT_ANALYST_MODEL_NAME=qwen3-max-preview
AGENT_FUNDAMENTALS_ANALYST_MODEL_NAME=deepseek-chat
AGENT_TECHNICAL_ANALYST_MODEL_NAME=glm-4-plus
AGENT_VALUATION_ANALYST_MODEL_NAME=moonshot-v1-32k
