🧠 SYNAPSE

A Modular AI Assistant with Tools, Memory, and Real-Time Streaming

Built with LangGraph, LangChain, and Cohere

Features • Architecture • Installation • Usage • Roadmap

---

📖 Overview

SYNAPSE is a graph-based AI assistant that combines the power of LangGraph orchestration, tool-aware reasoning, and conversational memory into a sleek Streamlit interface. Built for developers who want to create intelligent, context-aware chatbots with minimal friction.

Why SYNAPSE?

• 🎯 Modular Design — Easy to extend with new tools and capabilities
• 🔄 Graph-Based Reasoning — Sophisticated agent orchestration using LangGraph
• 💾 Conversational Memory — Maintains context across multi-turn conversations
• ⚡ Real-Time Streaming — Token-by-token response generation
• 🛠️ Tool Integration — Web search, stock prices, and custom tools
• 🎨 Clean UI — Professional Streamlit interface with session management

---

✨ Features

🤖 AI Core Graph-based agent orchestration with LangGraph Tool-enabled LLM (ChatCohere) Conditional tool execution & routing In-memory conversational state management	🎨 User Experience Streamlit-based chat interface Multi-threaded conversation switching Session state management Token-level streaming responses
🛠️ Tools & Integrations Web search (DuckDuckGo) Stock price lookup Structured data retrieval Extensible tool architecture	💾 Memory & State LangGraph checkpoint system Cross-session memory persistence Thread-based conversation management Context-aware responses

---

🏗️ Architecture

SYNAPSE follows a clean, modular architecture that separates concerns for maximum flexibility:

graph TB
    A[Streamlit Frontend] -->|User Input| B[LangGraph StateGraph]
    B -->|Process| C[ChatCohere LLM]
    C -->|Decision| D{Tool Needed?}
    D -->|Yes| E[ToolNode]
    D -->|No| F[Direct Response]
    E -->|Execute| G[Tool Functions]
    G -->|Results| C
    C -->|Stream| A
    B ---|Memory| H[(Checkpoint Store)]
    
    style A fill:#ff6b6b
    style B fill:#4ecdc4
    style C fill:#45b7d1
    style E fill:#96ceb4
    style H fill:#ffeaa7

Loading

Component Breakdown

Component	File	Responsibility
Backend	Backend.py	LangGraph state management, LLM orchestration, tool routing
Frontend	Frontend.py	Streamlit UI, conversation management, streaming display
Tools	Tools.py	Tool definitions, web search, stock prices, custom functions

---

📂 Project Structure

synapse-ai/
│
├── 📄 Backend.py          # Core agent logic & graph orchestration
├── 🎨 Frontend.py         # Streamlit chat interface
├── 🛠️ Tools.py            # Tool definitions & implementations
├── 📋 requirements.txt    # Python dependencies
├── 🔐 .env.example        # Environment variable template
└── 📖 README.md           # Project documentation

---

⚙️ Setup

Prerequisites

• Python 3.8+
• Cohere API Key (Get one here)

1️⃣ Clone the Repository

git clone https://github.com/Nabin68/SYNAPSE.git
cd SYNAPSE

2️⃣ Install Dependencies

pip install -r requirements.txt

Required packages:

langchain
langchain-cohere
langgraph
streamlit
python-dotenv
duckduckgo-search
yfinance

3️⃣ Configure Environment

Create a .env file in the project root:

COHERE_API_KEY=your_cohere_api_key_here

💡 Tip: Copy .env.example to .env and fill in your API key

---

🚀 Usage

Running the Application

streamlit run Frontend.py

The application will open in your default browser at http://localhost:8501

Using the Chat Interface

1. Start Chatting — Type your message in the input box
2. Watch Live Responses — See the AI respond in real-time with streaming
3. Switch Conversations — Use the sidebar to create or switch between threads
4. Tool Usage — The AI automatically uses tools when needed (search, stock prices, etc.)

Example Interactions

User: What's the current price of Apple stock?
🧠 SYNAPSE: [Uses stock price tool]
      The current price of AAPL is $195.83...

User: Search for the latest AI news
🧠 SYNAPSE: [Uses web search tool]
      Here are the latest developments in AI...

---

🛠️ Extending SYNAPSE

Adding New Tools

Create a new tool in Tools.py:

@tool
def my_custom_tool(query: str) -> str:
    """Description of what your tool does"""
    # Your implementation
    return result

Register it in the tool list:

tools = [web_search, stock_price, my_custom_tool]

Customizing the LLM

Modify Backend.py to use a different model or provider:

from langchain_openai import ChatOpenAI

model = ChatOpenAI(model="gpt-4", temperature=0.7)

---

🔮 Roadmap

Coming Soon

• 💾 SQLite-based persistent memory — Long-term conversation storage
• 📚 RAG integration — Document Q&A and knowledge retrieval
• 🔌 MCP server support — Model Context Protocol integration
• 🧰 Extended tool library — Weather, news, calculations, and more
• 🎨 Theme customization — Dark mode and custom color schemes
• 📊 Analytics dashboard — Conversation insights and usage stats
• 🌐 Multi-language support — Internationalization
• 🔒 Authentication — User accounts and access control

---

🤝 Contributing

Contributions are welcome! Here's how you can help:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

Development Guidelines

• Follow PEP 8 style guidelines
• Add docstrings to new functions
• Update README for new features
• Test thoroughly before submitting

---

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

---

👤 Author

Nabin

• GitHub: @Nabin68
• LinkedIn: Nabin Rouniyar

---

🙏 Acknowledgments

• LangChain — For the incredible LLM framework
• LangGraph — For graph-based agent orchestration
• Cohere — For powerful language models
• Streamlit — For the beautiful UI framework

---

📞 Support

If you encounter any issues or have questions:

• 🐛 Open an issue
• 💬 Start a discussion
• 📧 Email: nabingupta68@gmail.com

---

⭐ Star this repo if you find it helpful!

Made with ❤️ by Nabin