AccessPlus

AccessPlus is an agentic AI virtual assistant for banking. It leverages multiple AI agents, retrieval-augmented generation (RAG), and LLMs to provide secure, intelligent, and context-aware banking support via chat.

---

Features

• Multi-agent architecture (account, info, supervisor router)
• Retrieval-Augmented Generation (RAG) with CSV/FAISS knowledge bases
• LLM-powered natural language understanding and response generation
• Secure banking tools (get_my_balance, get_my_transactions, get_bank_policies)
• Stateful conversations with memory
• FastAPI/Flask web interface for real-time chat
• AI safety and observability

---

Architecture

User Chat UI
      |
Supervisor Router
  |           |
Account      Info
 Agent       Agent
  |           |
Banking Tools / Knowledge Base
      |
      LLM
      |
 Response Back

---

Installation

git clone https://github.com/TamerDotWork/AccessPlus.git
cd AccessPlus
python -m venv venv
source venv/bin/activate  # Linux/macOS
venv\Scripts\activate     # Windows
pip install -r requirements.txt
python main.py

---

Usage

from accessplus import AccessPlus

assistant = AccessPlus()
response = assistant.chat("Show my last 5 transactions")
print(response)

Expected Output:

1. - $50 Starbucks
2. - $120 Amazon
3. - $200 Salary Deposit
4. - $15 Netflix
5. - $30 Grocery Store

---

Extending AccessPlus

• Add new agents and register them in the supervisor router
• Integrate additional tools or APIs
• Swap LLMs (Gemini-Flash, Llama, Mistral, etc.)

---

License

MIT License