Installation

Set up your Python environment, install all dependencies, configure TA-Lib, and add your LLM API key.

Requirements

Before you begin, make sure you have conda installed. The following Python packages are installed via requirements.txt:

Package	Purpose
flask	Web interface server
yfinance	Real-time market data from Yahoo Finance
pandas, numpy	Data processing
matplotlib, mplfinance	Chart rendering
scipy	Signal processing utilities
TA-Lib	Technical indicator computation
langchain, langchain-core	LLM abstraction layer
langchain-openai	OpenAI provider integration
langchain-anthropic	Anthropic provider integration
langchain-qwq	Qwen provider integration
langgraph	Multi-agent graph orchestration
openai, anthropic	Provider SDK clients
Pillow	Image processing for chart analysis
requests, ipython	HTTP utilities and notebook support

Installation steps

1. Create and activate a conda environment

   Python 3.11 is required.

   conda create -n quantagents python=3.11
   conda activate quantagents
   
2. Install Python dependencies

   pip install -r requirements.txt
   
3. Handle TA-Lib installation issues (if needed)

   TA-Lib requires a C library that pip sometimes cannot build automatically. If the pip install fails, use the conda-forge package instead:

   conda install -c conda-forge ta-lib
   

   If you still encounter issues, visit the TA-Lib Python repository for platform-specific installation instructions covering Windows, macOS, and Linux.

4. Set your LLM API key

   Configure the API key for your chosen provider. See the API key setup section below for all three providers.

API key setup

QuantAgent supports OpenAI, Anthropic, and Qwen. Set the environment variable for your chosen provider, or enter it later through the web interface settings panel.

OpenAI

Get your API key from platform.openai.com/api-keys.

export OPENAI_API_KEY="your_openai_api_key_here"

Default models when using OpenAI:

• Agent LLM: gpt-4o-mini
• Graph LLM: gpt-4o

You can override these in default_config.py:

DEFAULT_CONFIG = {
    "agent_llm_model": "gpt-4o-mini",
    "graph_llm_model": "gpt-4o",
    "agent_llm_provider": "openai",
    "graph_llm_provider": "openai",
    "agent_llm_temperature": 0.1,
    "graph_llm_temperature": 0.1,
    "api_key": "sk-",
}

Anthropic

Get your API key from console.anthropic.com.

export ANTHROPIC_API_KEY="your_anthropic_api_key_here"

Default models when using Anthropic:

• Agent LLM: claude-haiku-4-5-20251001
• Graph LLM: claude-haiku-4-5-20251001

To switch the active provider at runtime via the Python API:

from trading_graph import TradingGraph

config = {
    "agent_llm_provider": "anthropic",
    "graph_llm_provider": "anthropic",
    "agent_llm_model": "claude-haiku-4-5-20251001",
    "graph_llm_model": "claude-haiku-4-5-20251001",
    "agent_llm_temperature": 0.1,
    "graph_llm_temperature": 0.1,
    "anthropic_api_key": "your_key_here",
}
trading_graph = TradingGraph(config=config)

Qwen (DashScope)

Get your API key from dashscope.console.aliyun.com. Note that DashScope is based in Singapore — you may experience higher latency depending on your location.

export DASHSCOPE_API_KEY="your_dashscope_api_key_here"

Default models when using Qwen:

• Agent LLM: qwen3-max
• Graph LLM: qwen3-vl-plus

To switch the active provider at runtime via the Python API:

from trading_graph import TradingGraph

config = {
    "agent_llm_provider": "qwen",
    "graph_llm_provider": "qwen",
    "agent_llm_model": "qwen3-max",
    "graph_llm_model": "qwen3-vl-plus",
    "agent_llm_temperature": 0.1,
    "graph_llm_temperature": 0.1,
    "qwen_api_key": "your_key_here",
}
trading_graph = TradingGraph(config=config)

All three providers require a vision-capable model. The Pattern and Trend agents generate K-line charts and pass them directly to the LLM for interpretation. Text-only models will cause analysis failures.

Verify your installation

Start the web interface to confirm everything is working:

python web_interface.py

Open http://127.0.0.1:5000 in your browser. If the page loads and the settings panel shows your API key as set, your installation is complete.

Troubleshooting

TA-Lib build failure on pip install Use conda install -c conda-forge ta-lib as described above, or follow the platform-specific instructions in the official repository.

API key not recognized Verify the environment variable is exported in the same shell session where you run python web_interface.py. Alternatively, enter the key directly in the web interface settings panel and click Save.

Data not available for a symbol QuantAgent uses Yahoo Finance via yfinance. Some symbols may have limited historical data or may not be available. Try a different date range or a standard symbol such as BTC, SPX, or AAPL.

Rate limit or billing errors Check that your API account has sufficient credits and that you are not exceeding the provider’s rate limits. Wait a moment and retry the analysis.