Agentic AI Chrome Extension — Multi-Step Stock Research with MCP

Описание: This demo shows how the Agentic Ticker Research backend powers a Chrome Extension that performs multi-step, reasoning-based stock analysis using Gemini 2.0 Flash, FastAPI, and MCP tools.

The system mimics a financial analyst through a four-layer cognitive architecture (Perception, Memory, Decision, and Action). It fetches real-time stock data, links recent news headlines, summarizes insights, and returns a concise reasoning-based analysis.

Key features:
• Verified system prompt using Gemini prompt evaluator
• MCP tools for stock, news, and summarization
• Pydantic-typed cognitive layers
• Chrome extension integration for real-time research

Built with: Python 3.13+, FastAPI, Gemini 2.0 Flash, MCP, and uv package manager.

GitHub: github.com/sushant097/Agentic-Architecture-EAG-Session6-Assignment