🤖 AI Copilot Assistant

An AI-powered personal copilot designed to bring conversational AI, intelligent agents, web search, document understanding and multiple AI workflows into a single web application.

---

🚀 Features

• 💬 AI Chat Assistant – Ask questions and receive AI-generated responses.
• 🔎 Web Search – Search the web for relevant information using DuckDuckGo.
• 🕒 Current Time Tool – Provides the current date and time through an AI tool.
• 📄 PDF Upload & Summarization – Upload PDF documents and work with their content.
• 🧠 PDF Reader / RAG Mode – Retrieve relevant information from uploaded documents.
• 🤖 Autonomous Agent Mode – Allows the AI agent to select appropriate tools for a task.
• 👥 CrewAI Research Crew – Uses multiple AI agents for research-oriented workflows.
• ☁️ Cloud LLM Support – Supports cloud-based LLM access through OpenRouter.
• 💻 Local LLM Support – Supports local language models through Ollama.
• 🎛️ Multiple Brain Modes – Different AI modes for different types of tasks.

---

🏗️ Project Architecture

AI-Copilot-Assistant/
│
├── backend/
│   ├── __init__.py
│   ├── agent.py
│   ├── crew.py
│   ├── engine.py
│   ├── llm_config.py
│   └── main.py
│
├── frontend/
│   ├── src/
│   │   ├── App.jsx
│   │   ├── firebase.js
│   │   ├── index.css
│   │   └── main.jsx
│   ├── index.html
│   ├── package.json
│   ├── package-lock.json
│   ├── postcss.config.js
│   ├── tailwind.config.js
│   └── vite.config.js
│
├── Docs/
│   └── crewAIDocs.txt
│
├── .env.example
├── .gitignore
├── GUIDE.md
├── README.md
├── agent-demo.png
├── package-lock.json
├── requirements.txt
├── setup.sh
└── visualize_db.py

---

🛠️ Tech Stack

Backend

• Python
• FastAPI
• LangChain
• LangChain Community
• CrewAI

AI / LLM

• Ollama
• Gemma
• OpenRouter
• LangChain Agents

Tools & Retrieval

• DuckDuckGo Search
• PDF Processing
• Retrieval-Augmented Generation (RAG)
• Chroma Vector Database

Frontend

• React.js
• JavaScript
• HTML
• CSS
• Vite
• Node.js
• Firebase

---

🧠 AI Capabilities

PDF Reader / RAG

The application can process uploaded PDF documents and retrieve relevant information to answer questions based on their content.

Upload PDF
     ↓
Process Document
     ↓
Create Embeddings
     ↓
Store in Vector Database
     ↓
Retrieve Relevant Content
     ↓
Generate AI Response

---

Autonomous Agent

The Autonomous Agent can select and use available tools based on the user's query.

User Query
    ↓
AI Agent
    ↓
Select Appropriate Tool
    ├── Web Search
    └── Current Time
    ↓
Generate Response

---

Research Crew

The Research Crew uses multiple AI agents to perform research-oriented tasks and combine their outputs into a final response.

---

🤖 Local & Cloud LLM Support

The application supports both local and cloud-based language models.

Local LLM

Ollama can be used to run supported models locally.

Application
     ↓
LangChain
     ↓
Ollama
     ↓
Local Gemma Model

Cloud LLM

OpenRouter provides access to cloud-based models through an OpenAI-compatible API.

Application
     ↓
LangChain / OpenAI-compatible API
     ↓
OpenRouter
     ↓
Selected LLM

---

⚙️ Installation

1. Clone the Repository

git clone https://github.com/anmolshr56/AI-Copilot-Assistant.git
cd AI-Copilot-Assistant

2. Create a Python Virtual Environment

Windows:

python -m venv venv
venv\Scripts\activate

3. Install Python Dependencies

pip install -r requirements.txt

4. Install Frontend Dependencies

cd frontend
npm install
cd ..

---

🔐 Environment Variables

Create a .env file in the project root.

Example:

OPENROUTER_API_KEY=your_openrouter_api_key

Never upload your real API key or other secrets to GitHub.

The .env file should remain ignored by Git.

---

🦙 Ollama Setup

For local LLM support, install Ollama and download the required model.

Example:

ollama pull gemma2:2b

Make sure Ollama is running before using local model functionality.

---

▶️ Running the Application

Start the Backend

From the project root:

python -m backend.main

Start the Frontend

Open another terminal:

cd frontend
npm run dev

Open the local URL shown by the frontend development server.

---

🔄 How It Works

                ┌───────────────────┐
                │       User        │
                └─────────┬─────────┘
                          │
                          ▼
                ┌───────────────────┐
                │   AI Copilot UI   │
                └─────────┬─────────┘
                          │
                          ▼
                ┌───────────────────┐
                │    Backend API    │
                └─────────┬─────────┘
                          │
            ┌─────────────┼─────────────┐
            │             │             │
            ▼             ▼             ▼
       ┌─────────┐   ┌──────────┐   ┌──────────┐
       │ Ollama  │   │OpenRouter│   │  Agents  │
       └─────────┘   └──────────┘   └────┬─────┘
                                         │
                             ┌───────────┴───────────┐
                             │                       │
                             ▼                       ▼
                        Web Search              PDF / RAG

---

🎯 Key Learning Areas

This project demonstrates practical implementation of:

• LLM-powered application development
• LangChain workflows
• AI agent architecture
• Tool calling
• Web search integration
• Retrieval-Augmented Generation (RAG)
• PDF document processing
• Local LLM deployment with Ollama
• Cloud LLM integration using OpenRouter
• Multi-agent workflows using CrewAI
• Backend and frontend integration

---

🔮 Future Improvements

• 🔐 User authentication
• 💾 Persistent chat history
• 🎙️ Voice input and output
• 📚 Support for additional document formats
• ⚡ Streaming AI responses
• 🧩 More specialized AI agents
• 🌐 Production deployment
• 📊 Usage analytics

---

👨‍💻 Author

Anmol Sharma

B.Tech CSE Student

GitHub: https://github.com/anmolshr56

---

📸 Project Demo

Autonomous AI Agent with Web Search

The AI Copilot can operate in autonomous agent mode and perform web searches to retrieve up-to-date information.