Stock Analysis Chatbot

This project implements a sophisticated stock analysis chatbot designed to provide comprehensive insights and recommendations for stock tickers or companies. Leveraging a suite of specialized tools, the chatbot gathers real-time market data, fundamental company information, recent news, and trend analysis to generate a concise, evidence-based sentiment.

Features

• Real-time Market Data (price_ohlcv): Fetches the most recent market price and OHLCV (Open, High, Low, Close, Volume) data. It prioritizes intraday data (last 10 minutes/candles) and falls back to daily candles if intraday is unavailable.
• Fundamental Analysis (fundamentals): Gathers rich fundamental data from Yahoo Finance, including company information (sector, industry), valuation metrics (P/E, market cap), profitability (margins, ROE), growth rates, financial health (debt, cash), and more.
• News Summarization (new_summariser): Collects fresh, Google-trending news articles related to the specified stock using Serper.dev. It then downloads the full text of these articles and uses Groq's advanced LLM (llama-3.3-70b-versatile) to generate a concise, investment-relevant summary, highlighting price-moving events, numerical highlights, sentiment signals, and risks/catalysts.
• Trend Analysis (trend_analysis): Performs technical trend analysis using 7-day or 30-day hourly OHLCV data. It calculates various technical indicators such as EMA crossovers, MACD, RSI, Bollinger Bands, and volume trends to provide a trend verdict and confidence score.
• AI-Powered Recommendations: Synthesizes outputs from all tools into a final recommendation with a sentiment (bullish, bearish, neutral), confidence score, detailed rationale, and key signals in a structured JSON format.

Architecture

The chatbot is built using langgraph for managing the conversational flow and tool orchestration.

• StateGraph: Defines the graph structure, connecting different nodes.
• Chatbot Node: An LLM node that integrates with Google Generative AI (gemini-2.5-pro) and binds all financial tools. It's responsible for processing user input, calling tools, and generating the final response.
• ToolNode: A prebuilt langgraph node that executes the defined financial analysis tools (price_ohlcv, fundamentals, new_summariser, trend_analysis).
• Router Node: Directs the graph flow. If the LLM response includes tool calls, it routes to the tools node; otherwise, it ends the graph.

Setup and Installation

Prerequisites

• Python 3.9+
• uv package manager (used for dependency management and virtual environments)
• API keys for:
  • Google Generative AI (for ChatGoogleGenerativeAI)
  • Serper.dev (for stock_news in tools/news.py)
  • Groq (for summarize_with_groq in tools/news.py)

Environment Variables

Create a .env file in the root directory and add your API keys:

GOOGLE_API_KEY=your_google_generative_ai_api_key
SERPER_API_KEY=your_serper_dev_api_key
GROQ_API_KEY=your_groq_api_key

Installation

This project uses uv for dependency management, a fast Python package installer and resolver.

1. Clone the repository:

   git clone https://github.com/Dharshan2208/stock-analysis-chatbot.git
   cd stock-analysis-chatbot

2. Install uv (if you don't have it):

   pip install uv

Running the Chatbot

To start the stock analyst chatbot:

uv run main.py

You can then interact with the chatbot in your terminal, providing stock tickers (e.g., 'AAPL', 'MSFT', 'TSLA') or questions to receive financial recommendations.

Tool Details

tools/market.py

Contains the price_ohlcv tool:

• price_ohlcv(symbol: str): Fetches current price and OHLCV data using Yahoo Finance. Prioritizes 1-minute intraday data, falling back to 1-day historical data.

tools/fundamentals.py

Contains the fundamentals tool:

• fundamentals(symbol: str): Retrieves extensive fundamental data for a given stock symbol from Yahoo Finance, including valuation, profitability, growth, and financial health metrics.

tools/news.py

Contains the stock_news and new_summariser tools:

• stock_news(symbol: str, recency: str = "past 24 hours"): Fetches recent news articles for a stock using Serper.dev (Google News search).
• new_summariser(symbol: str): Orchestrates news fetching, full article text extraction (using newspaper3k), and summarization (using Groq's llama-3.3-70b-versatile LLM). It provides a detailed, investment-focused summary.

tools/trend.py

Contains price_ohlcv_month_hourly, price_ohlcv_weekly_hourly, and trend_analysis tools:

• price_ohlcv_month_hourly(symbol: str): Retrieves approximately 30 days of 30-minute OHLCV data for trend detection.
• price_ohlcv_weekly_hourly(symbol: str): Retrieves 7 days of 1-hour OHLCV data for shorter-term trend analysis.
• trend_analysis(symbol: str, period: str = "7d"): The core trend analysis engine. It processes hourly OHLCV data (7-day or 30-day based on period) to calculate various technical indicators (EMA, MACD, RSI, Bollinger Bands, Volume) and provides a trend verdict, confidence, and key signals.

Example Outputs

Below are two example output screenshots :

Note

• This is a personal project, and I'm actively improving it as I learn more about AI agents, and financial analysis.
• New features, better pipelines, and more polished tooling will be added over time.
• Expect frequent updates to structure, tools, and agent reasoning.
• Feel free to customize and expand any part of this system depending on your future goals.

Contribution

Feel free to fork the repository, open issues, or submit pull requests.