A two-agent CrewAI application that retrieves Yahoo Finance market data, produces an analyst summary, and makes a Buy/Sell/Hold recommendation based on that summary.
- Create and activate a Python virtual environment.
- Install dependencies with
pip install -r requirements.txt. - Copy
env_template.txtto.envand setOPENAI_API_KEY.
Analyze one or several exchange ticker symbols:
python main.py AAPL TSLA CAP.PARunning python main.py without arguments prompts you to enter a ticker. This is convenient for IDE run configurations because you can choose a different ticker on every run.
Use Yahoo Finance ticker notation (for example, CAP.PA for Capgemini in Paris). Market data may be delayed. Output is for informational purposes only and is not financial advice.
Start the graphical interface from the project directory:
streamlit run streamlit_app.pyIf the streamlit command is not available directly, use:
python -m streamlit run streamlit_app.pyEnter a Yahoo Finance ticker in the form and select Analyze stock. The analysis can take a little while because it runs both CrewAI agents.
The GUI lets you choose OpenAI or a local Ollama server. For OpenAI, use the key from .env or enter one in the password field. For Ollama, the default server is http://192.168.1.168:11434/; available models are loaded from /api/tags, with a manual model-name fallback if discovery is unavailable.