Sammy - Personal Finance Agent 💰

Sammy is a multi-agent conversational AI system designed to democratize financial literacy. It provides personalized financial education, portfolio analysis, and goal planning using real-time market data and a curated knowledge base.

🚀 Key Features

• Multi-Agent Orchestration: Powered by CrewAI with a hierarchical (Manager-led) architecture.
• 6 Specialized Agents:
  • Finance Q&A Agent: General education via Pinecone RAG.
  • Portfolio Analysis Agent: Reviews diversification and risk in your holdings.
  • Market Analysis Agent: Real-time stock quotes via Alpha Vantage.
  • Goal Planning Agent: Retirement and savings goal calculations.
  • News Synthesizer Agent: Contextualizes market news via Yahoo Finance.
  • Tax Education Agent: Explains complex tax concepts and account types.
• 3-Tier LLM Resilience: Automated fallback sequence: Gemini 2.5 Flash → OpenAI GPT-4o → Local Llama 3.1 (Ollama).
• RAG Knowledge Base: 110+ curated articles indexed in a high-dimensional Pinecone vector store.
• Observability: Full tracing and monitoring integrated with LangSmith.

🛠️ Project Structure

• src/agents/: Specialized agent definitions (managed via crew_setup.py).
• src/core/: Core logic and shared base classes.
• src/data/: Educational articles, glossary, and sample portfolio.
• src/rag/: Pinecone indexer and search logic.
• src/web_app/: Streamlit dashboard interface.
• src/utils/: Custom tools for market data, news, and goal planning.
• src/workflow/: CrewAI hierarchical orchestration.

⚙️ Setup Instructions

1. Environment Setup:

   python -m venv venv
   source venv/bin/activate
   pip install -r requirements.txt

2. API Configuration: Create a .env file based on .env.example:

   • GOOGLE_API_KEY: For Gemini and Embeddings.
   • ALPHA_VANTAGE_API_KEY: For real-time market data.
   • PINECONE_API_KEY & PINECONE_INDEX_NAME: For the vector store.
   • LANGCHAIN_API_KEY (Optional): For LangSmith tracing.

3. Initialize Knowledge Base: If the Pinecone index is not yet populated:

   python3 src/rag/indexer.py

4. Run the Application:

   streamlit run src/web_app/main.py

🌐 Deployment

Sammy is designed to be hosted on Streamlit Community Cloud and integrated into your portfolio at samiraryamane.com.

1. Push your code to a GitHub repository.
2. Connect the repository to Streamlit Cloud.
3. Add your environment variables to the "Secrets" dashboard in Streamlit Cloud.
4. Link the deployment URL to your portfolio website.

---

Disclaimer: For educational purposes only. Not financial advice.