A smart crypto analytics app powered by Phi Agents, OpenAI, and Streamlit that answers crypto-related questions using real-time data, market fundamentals, and live web searches. This multi-agent system can analyze project descriptions, forecast 2025 trends, and tell you whether it's worth investing — all in one click.
- 🧠 Uses a multi-agent system to fetch insights from both web and financial sources
- 📊 Presents fundamentals and analyst recommendations using tables
- 🔎 Real-time live crypto prices from web searches (2025 outlook)
- 📢 Includes Buy or Not recommendations with reasons
- 🧾 Answers are structured in Markdown, often with sources
- Frontend/UI: Streamlit
- Agents Framework: Phi
- AI Models: OpenAI (via
openaiPython SDK) - Data Tools:
phi.tools.yfinancefor crypto fundamentals and analyst recommendationsphi.tools.duckduckgofor live web data
git clone <repository-url>
cd <repository-folder>pip install -r requirements.txtCreate a .env file in the root directory and add your OpenAI API key:
OPENAI_API_KEY=your_openai_api_key_herestreamlit run research.py-
User enters a crypto-related query like:
"Bitcoin Price Prediction"or"Should I invest in Ethereum?" -
Two AI agents are activated:
- 📊 YFinance Agent: Analyzes project fundamentals
- 🌐 Web Search Agent: Searches web for real-time price & sentiment
-
Multi-Agent Coordinator merges responses, analyzes and recommends
-
Output includes:
- Project background
- Price outlook
- Fundamental analysis
- Recommendation (Buy/Hold/Sell)
- Sources of information
- ✅ "Bitcoin price prediction for 2025"
- ✅ "Is Solana a good investment now?"
- ✅ "Dogecoin fundamentals and analyst recommendations"
- ✅ "Latest updates on Polygon project"
- Python 3.8+
- OpenAI API key
- Internet connection for real-time DuckDuckGo search
- If no response appears, ensure:
- You have an active OpenAI API key
- DuckDuckGo search doesn’t fail due to rate limits
- You are using a valid crypto keyword
MIT License