This repository showcases my hands-on projects while learning about AI Agents using LangChain, Ollama, and Python. It includes both minimal setups and advanced interactive applications that demonstrate real-world agent use cases like memory handling, voice interaction, and local LLM inference.

.
├── basic_ai_agent/ # Basic text agent with and without memory
│ ├── basic_ai_agent.py
│ ├── basic_ai_agent_with_memory.py
│ ├── streamlit_app.py
│ ├── requirements.txt
│ └── README.md
│
├── pdf_summary_bot/ # QA-based PDF summarizer agent
│ ├── app_basic_qa.py
│ ├── app_summary_qa.py
│ ├── requirements.txt
│ └── README.md
│
├── voice_assistant/ # Voice assistant with terminal + Streamlit UI
│ ├── ai_voice_assistant_and_streamlit_app.py
│ ├── streamlit_app.py
│ ├── requirements.txt
│ └── README.md
│
├── web_scraper_agent/ # Agent that scrapes data using tools + memory
│ ├── ai_web_scraper.py
│ ├── ai_web_scrapper_faiss.py
│ ├── requirements.txt
│ └── README.md
│
├── .gitignore
├── README.md # ← You are here
└── requirements.txt # Master requirements for full repo

Minimal examples to:

• Load local LLMs (like phi, mistral, etc.) via Ollama
• Run single-turn and memory-supported agents
• Try a web UI version with Streamlit

Ask questions or summarize based on uploaded PDF content:

• app_basic_qa.py: Basic question answering
• app_summary_qa.py: Summary-focused version

Talk to your LLM with:

• A terminal-based CLI voice chat
• A full Streamlit-based voice chat UI

Give queries like "Find top Python tutorials" and get results scraped live:

• ai_web_scraper.py: Basic tool-based agent
• ai_web_scrapper_faiss.py: Uses memory and FAISS for smarter recall

1. Clone this repo

   git clone https://github.com/your-username/ai-agent-projects.git
   cd ai-agent-projects

2. Install Dependencies

   pip install -r requirements.txt

3. Make sure Ollama is running

   ollama run phi

4. Run a project

   cd voice_assistant
   python ai_voice_assistant_and_streamlit_app.py
   # OR
   streamlit run streamlit_app.py

Each folder includes its own requirements.txt, but a master file is available at root if you want to install everything together.

• Python 3.9+
• Ollama running locally with a model like phi, mistral, or llama2
• Microphone + speaker (for voice apps)

• LangChain for the powerful agent ecosystem
• Ollama for local, private LLM serving
• Streamlit for interactive UIs

This repo is for educational purposes. MIT License. Feel free to fork, use, and modify with attribution.

• Add file upload / PDF reading support
• Build a full research assistant
• Experiment with tools, memory, and agents

Pull requests and suggestions are welcome!