AI Trader is a full-stack application that uses a multi-agent system to perform different financial analyses about a specific ticker symbol served by a backend system using FastAPI.
The system is built using LangGraph and LLMs. The graph has three main nodes:, the fundamental analysis node, the technical analysis node, and the manager node.
Fundamental analysis node is a single agent that researches and provides reports about the fundamentals of the financial asset.
Technical analysis node is built with two agents; one is responsible for doing research about technical analysis, defining which chart configurations should be used, and the second agent analyzes the chart image to provide the final report.
Finally, the manager receives the reports from both sectors and provides requests from the sectors if needed; otherwise, they provide a final report.
The backend app integrates an automated deployment to Google Cloud with GitHub Actions CI/CD. It builds a Dockerfile image at Cloud Build and deploys the image to Cloud Run.
The frontend app includes a simple Flutter mobile application that sends requests to the AITrader API and displays the responses at a friendly interface for the user.
This application is for learning purposes only; no personal investment advice should be taken.