Stock Analysis Telegram Agent 📈

A multi-agent autonomous trading analysis system that provides real-time stock recommendations via Telegram. Built with AutoGen, TwelveData, and NewsAPI.

🚀 Overview

This agent leverages a "Chain of Thought" approach to stock analysis. When a user sends a stock symbol (e.g., AAPL, NVDA), the system orchestrates multiple agents to fetch technical data and market news, perform sentiment analysis, and generate a unified trade recommendation.

🧠 Program Logic & Architecture

The system follows a structured pipeline designed to mimic a professional day trader's workflow:

1. Parallel Data Fetching

Upon receiving a symbol, the system triggers parallel requests to:

• TwelveData API: Fetches OHLCV (Open, High, Low, Close, Volume) data for three timeframes: 1m, 15m, and 1h.
• NewsAPI: Fetches the most recent and relevant news articles from the past 24 hours.

2. Sentiment Analysis (SentimentAnalyzer Agent)

The news articles are passed to the SentimentAnalyzer agent. This agent:

• Evaluates market reaction and news impact.
• Determines a category (Positive, Neutral, Negative).
• Calculates a sentiment score (-1 to 1).
• Provides a rationale with key headlines.

3. Trade Recommendation (TradeDecisionAgent)

The TradeDecisionAgent aggregates the multi-timeframe candle data and the sentiment report. It follows these steps:

• Trend Confirmation: Uses 1h data to identify the primary trend.
• Entry Signals: Analyzes 1m and 15m action for optimal entry levels.
• Sentiment Integration: Weighs the technical signals against real-time market sentiment.
• Output Generation: Produces a structured recommendation including Buy/Sell/Hold, Entry Price, Stop-Loss, Target Price, and a detailed Rationale.

📦 Installation (using uv)

This project uses uv, an extremely fast Python package manager.

1. Install uv

If you haven't installed uv yet, run:

# macOS/Linux
curl -lsSf https://astral.sh/uv/install.sh | sh

2. Clone the Repository

git clone https://github.com/alfredang/stock_analysis_telegram_agent.git
cd stock_analysis_telegram_agent

3. Setup Environment

Initialize the project and install dependencies:

uv sync

This will automatically create a virtual environment in .venv and install all required packages.

4. Configuration

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

OPENAI_API_KEY=your_openai_key
TWELVEDATA_API_KEY=your_twelvedata_key
NEWSAPI_API_KEY=your_newsapi_key
TELEGRAM_BOT_TOKEN=your_telegram_bot_token

🛠 Usage

Run the Telegram Bot

uv run stock_agent.py

Run the Logic Test Script (Interactive)

To verify the analysis logic without Telegram, you can run the interactive test script:

uv run test_analysis.py

When prompted, enter a stock ticker (e.g., AAPL, NVDA). The script will fetch the multi-timeframe candle data, perform sentiment analysis, and display the final unified trade recommendation with rationale directly in your terminal.

📝 License

MIT