🤖 Agentic AI and Generative AI Practice

📘 Introduction

Welcome to the Agentic AI and Generative AI Practice repository – a hands-on collection of projects and experiments exploring the cutting edge of Agentic AI and Generative AI. From multi-agent orchestration to tool‑augmented reasoning, this repo covers a wide spectrum of frameworks and platforms including CrewAI, LangChain/LangGraph/LangSmith, LlamaIndex, Smolagents, Agno/Phidata, Cerebras, LiveKit, MCP servers, and Local LLM Inference Tools.

Whether you're building semantic classifiers, real‑time sales agents, research assistants, or RAG pipelines, you'll find practical, runnable examples with detailed instructions and screenshots. The repository also includes foundational course materials from LangChain Academy and the Hugging Face Agents Course.

📑 Table of Contents

🤖 Agentic AI and Generative AI Practice

⚙️ Technical Stack

The projects leverage a diverse set of modern AI and developer tools:

Category	Technologies
Agent Frameworks	CrewAI, LangChain, LangGraph, LangSmith, LlamaIndex, Smolagents, Agno, Phidata, LangBase
LLM Providers & Hardware	Cerebras, Gemini, Groq, OpenAI, Ollama, vLLM, Llamacpp, LM Studio, Jan, Koboldcpp, OpenClaw
Voice & Real‑time	LiveKit, Cartesia, DeepGram, ElevenLabs
Search & Data	Exa, Neo4j, SQL, Chroma, RAG pipelines
Development & Deployment	Python, TypeScript, Streamlit, FastAPI, MCP, Docker, UV, Conda
Observability	LangSmith, LangFuse, Arize AI
Cloud & Notebooks	Google Colab, Jupyter, VS Code, Claude Desktop, GitHub Copilot

🏗️ Repository Structure

Agentic-AI-and-Generative-AI-Practice/
├── README.md
├── Agno_and_Phidata_Apps/
│   ├── agents.db
│   ├── 1_Semantic_Classifier_and_Sports_Coach_Agents_Agno.py
│   ├── 2_Tech_Research_and_Data_Analysis_Agents_Phidata.py
│   ├── Imgs/
│   │   ├── Agno_AgentOS_Agents.png
│   │   └── Phidata_PlaygroundUI_Agents.png
│   ├── requirements.txt
│   └── instructions.txt
├── Cerebras_Cartesia_LiveKit_Exa_DeepGram_Apps/
│   ├── Automated-User-Research-[Cerebras, LangGraph, LangSmith].ipynb
│   ├── Build-Your-Own-Perplexity-[Exa, Cerebras].ipynb
│   ├── Real-Time-Sales-Agent-[Livekit, Cartesia, Deepram, Cerebras].ipynb
│   ├── Imgs/
│   │   ├── Automated_User_Research.png
│   │   ├── Build_Your_Own_Perplexity.png
│   │   └── Real_Time_Sales_Agent.png
│   ├── requirements.txt
│   └── instructions.txt
├── CrewAI_Apps/
│   ├── pyproject.toml
│   ├── restricted_func.py
│   ├── 1_Email_Agent_with_Tools.ipynb
│   ├── 2_Research_Agent_with_Tools.ipynb
│   ├── 3_Marketing_Agent_with_Config.py
│   ├── 4_Legalising_Agent_with_Config.py
│   ├── anaconda_projects/
│   ├── config_legal_agents/
│   │   ├── agents.yaml
│   │   └── tasks.yaml
│   ├── config_market_agents/
│   │   ├── agents.yaml
│   │   └── tasks.yaml
│   ├── resources/
│   │   ├── for_3/
│   │   └── for_4/
│   ├── Imgs/
│   │   ├── CrewAI_Email_Agent.png
│   │   ├── CrewAI_Legal_Agent.png
│   │   ├── CrewAI_Marketing_Agent.png
│   │   └── CrewAI_Research_Agent.png
│   ├── requirements.txt
│   └── instructions.txt
├── LangBase_Apps/
│   ├── package.json
│   ├── package-lock.json
│   ├── agents.ts
│   ├── create-memory.ts
│   ├── create-pipe.ts
│   ├── index.ts
│   ├── upload-docs.ts
│   ├── docs/
│   │   └── statistical-concepts.txt
│   ├── node_modules/
│   ├── Imgs/
│   │   ├── LangBase_Agent_CLI.png
│   │   └── LangBase_Agent_GUI.png
│   ├── requirements.txt (if any)
│   └── instructions.txt
├── LangChain_LangGraph_LangSmith_Apps/
│   ├── 1)_LLM_Application_Langchain.ipynb
│   ├── 2)_Semantic_Search_Engine.ipynb
│   ├── 3)_Extraction_and_Classification.ipynb
│   ├── 4)_LangChain_Chatbot_along_with_LangGraph.ipynb
│   ├── 5)_End_to_End_Agent_in_LangChain.ipynb
│   ├── 6)_Retreival_Augmented_Generation_with_LangChain_Multiple_Ingestion.ipynb
│   ├── 7)_Retreival_Augmented_Generation_with_LangChain_through_LangChain_Docs.ipynb
│   ├── 8)_Retreival_Augmented_Generation_Summarizer_with_Stuff_and_MapReduce.ipynb
│   ├── 9)_Building_Question_Answer_System_over_SQLdb.ipynb
│   ├── 10)_Building_Question_Answer_System_over_a_Graph_Database.ipynb
│   ├── 11)_Building_Restaurant_Name_Generator_with_Gemini.ipynb
│   ├── Intro-to-LangChain/
│   ├── Intro-to-LangGraph/
│   ├── Intro-to-LangSmith/
│   ├── LangTrio_Agents/
│   ├── Notebook_Input_Data/
│   ├── Restaurant_Details_Generator/
│   │   └── frontend_streamlit.py
│   ├── Imgs/
│   │   ├── LangChain_Gemini_Restaurant_Generator_1.png
│   │   ├── LangChain_Gemini_Restaurant_Generator_2.png
│   │   ├── LangTrio_QA_Neo4J_Dataset.png
│   │   └── LangTrio_Retrieval_Augmented_Generator.png
│   ├── requirements.txt
│   └── instructions.txt
├── LlamaIndex_and_ArizeAI_Apps/
│   ├── agents.ipynb
│   ├── components.ipynb
│   ├── tools.ipynb
│   ├── workflows.ipynb
│   ├── Imgs/
│   │   ├── LlamaIndex_Agents.png
│   │   ├── LlamaIndex_ArizeAI_Components.png
│   │   ├── LlamaIndex_Tools.png
│   │   └── LlamaIndex_Workflows.png
│   ├── requirements.txt
│   └── instructions.txt
├── LocalLLM_Inference_Softwares/
│   ├── Commands_and_Scripts.md
│   ├── Modelfile
│   ├── 1_Ollama_API_Working.py
│   ├── 2_Custom_Knowitall_Model.py
│   ├── 3_LLM_Function_Access.py
│   ├── 4_Grocery_List_Categorizer.py
│   ├── 5_Streamlit_GUI_RAG.py
│   ├── 6_ElevenLabs_STT_RAG.py
│   ├── Input_Data/
│   ├── Application_Projects/
│   │   ├── AI-Recruiter-Agency/
│   │   ├── AI-Travel-Agents/
│   │   ├── News-Summarizer/
│   │   └── Ollama-Vision/
│   ├── Imgs/
│   │   ├── JanAI_App.png
│   │   ├── Koboldcpp_App.png
│   │   ├── Llamacpp_CLI.png
│   │   ├── LMStudio_App.png
│   │   ├── Ollama_App.png
│   │   ├── Ollama_CLI.png
│   │   ├── Ollama_OpenWebUI.png
│   │   ├── Openclaw_CLI.png
│   │   └── Vllm_CLI.png
│   ├── requirements.txt
│   └── instructions.txt
├── MCP_Server_Tools/
│   ├── main.py
│   ├── mcp.json
│   ├── sample_claude_desktop_config
│   ├── sample_mcp
│   ├── uv.lock
│   ├── pyproject.toml
│   ├── ExplanatoryVersion/
│   ├── MainCode/
│   │   ├── Scenario1/
│   │   ├── Scenario2/
│   │   ├── Scenario3/
│   │   └── deployment/
│   ├── .vscode/
│   │   └── mcp.json
│   ├── Imgs/
│   │   ├── Claude_Desktop_ToyDatasetMCP1.png
│   │   ├── Claude_Desktop_ToyDatasetMCP2.png
│   │   ├── Github_Copilot_CalculatorMCP.png
│   │   ├── Github_Copilot_FeedSearchMCP.png
│   │   ├── MCP_ServerInspect_Scenario1.png
│   │   ├── MCP_ServerInspect_Scenario2.png
│   │   └── MCP_ServerInspect_Scenario3.png
│   ├── requirements.txt
│   └── instructions.txt
├── Smolagent_and_LangFuse_Apps/
│   ├── code_agents.ipynb
│   ├── multiagent_notebook.ipynb
│   ├── retrieval_agents.ipynb
│   ├── tool_calling_agents.ipynb
│   ├── tools.ipynb
│   ├── vision_agents.ipynb
│   ├── Imgs/
│   │   ├── Smolagents_LangFuse_Code_Agent.png
│   │   ├── Smolagents_Multi_Agents_Notebook.png
│   │   ├── Smolagents_Retrieval_Agents.png
│   │   ├── Smolagents_Tool_calling_Agents.png
│   │   ├── Smolagents_Tools.png
│   │   └── Smolagent_Vision_Agents.png
│   ├── requirements.txt
│   └── instructions.txt
└── .gitignore

🚀 Project Showcase

Each folder contains self-contained projects with code, instructions, and screenshots. Below is a summary of what you'll find.

📁 Agno_and_Phidata_Apps

Technologies: Agno (AgentOS), Phidata, Semantic Search, Data Analysis
Goal: Build two distinct agents:

A semantic classifier & sports coach using Agno.
A tech research & data analysis duo using Phidata's playground UI.

Key Images:

Agno AgentOS Agents	Phidata Playground UI

Instructions:

Install dependencies: pip install -r requirements.txt
Run each script: python Agno_and_Phidata_Apps/1_Semantic_Classifier_and_Sports_Coach_Agents_Agno.py (and similarly for the second).

📁 Cerebras_Cartesia_LiveKit_Exa_DeepGram_Apps

Technologies: Cerebras, LangGraph, LangSmith, Exa, LiveKit, Cartesia, DeepGram, Google Colab
Goal: Three real‑world agentic applications:

Automated User Research – combines Cerebras speed with LangGraph orchestration.
Build Your Own Perplexity – search‑powered Q&A using Exa and Cerebras.
Real‑Time Sales Agent – voice‑enabled agent with LiveKit, Cartesia, DeepGram, and Cerebras.

Key Images:

Automated User Research	Build Your Own Perplexity	Real‑Time Sales Agent

Instructions:
Run the notebooks in Google Colab (recommended for free GPU). Each notebook is self‑contained and includes setup cells.

📁 CrewAI_Apps

Technologies: CrewAI, YAML configuration, Tools (e.g., restricted_func)
Goal: Explore CrewAI's multi‑agent patterns:

Email Agent with tools (Jupyter)
Research Agent with tools (Jupyter)
Marketing Agent with config (Python)
Legal Agent with config (Python)

Key Images:

Email Agent	Research Agent

Marketing Agent	Legal Agent

Instructions:

Create a virtual environment (Conda recommended) with Python version from .python-version.
Install dependencies: pip install -r requirements.txt
Run files: python CrewAI_Apps/3_Marketing_Agent_with_Config.py etc.
(Jupyter notebooks can be opened directly.)

📁 LangBase_Apps

Technologies: LangBase (TypeScript), Memory, Pipes, Document upload, CLI/GUI
Goal: Build a LangBase agent from scratch:

Create memory, upload documents, create pipes, and finally run the agent via CLI and GUI.

Key Images:

CLI Agent	GUI Agent

Instructions:

Ensure Node.js and npm are installed.
Inside the folder, run: npm install langbase dotenv
Execute the TypeScript files in order:
npx tsx create-memory.ts → upload-docs.ts → create-pipe.ts → index.ts

📁 LangChain_LangGraph_LangSmith_Apps

Technologies: LangChain, LangGraph, LangSmith, Gemini, Streamlit, Neo4j, SQL, RAG
Goal: A comprehensive suite covering everything from basic LLM apps to advanced RAG and QA over databases. Highlights:

11 Jupyter notebooks (LLM apps, semantic search, extraction, chatbots, RAG, SQL/graph QA)
Restaurant Name Generator with Gemini + Streamlit frontend
LangChain Academy tutorial folders (Intro-to-LangChain/Graph/Smith)

Key Images:

Restaurant Generator (1)	Restaurant Generator (2)

QA over Neo4j	RAG Summary

Instructions:

For the Restaurant Generator:
1. Add API keys to .env.
2. Create a virtual environment and install requirements.txt.
3. Run streamlit run Restaurant_Details_Generator/frontend_streamlit.py.
For other notebooks: follow steps 1‑2 above, then run the notebooks.
(The Intro-to-* folders contain course materials from LangChain Academy.)

📁 LlamaIndex_and_ArizeAI_Apps

Technologies: LlamaIndex, Arize AI, Agents, Tools, Workflows
Goal: Explore LlamaIndex's agentic capabilities with Arize for observability:

Agents, components, tools, and workflows notebooks.

Key Images:

Agents	Components

Tools	Workflows

Instructions:
Run the notebooks in Google Colab for optimal performance.

📁 LocalLLM_Inference_Softwares

Technologies: Ollama, vLLM, Llamacpp, LM Studio, Jan, Koboldcpp, OpenClaw, ElevenLabs, Streamlit
Goal: Hands‑on with local LLM inference engines:

Scripts to interact with Ollama (API, custom model, function calling, grocery categorizer)
Streamlit RAG app, ElevenLabs STT+RAG app
Commands & scripts for various local software (Jan, Koboldcpp, etc.)
Application projects: AI Recruiter, Travel Agents, News Summarizer, Ollama Vision

Key Images:

Ollama CLI	Ollama App	Ollama WebUI

Jan App	Koboldcpp	LM Studio

Openclaw CLI	Llamacpp CLI	vLLM CLI

Instructions:

Set up a virtual environment and install requirements.txt.
Add API keys to .env (if needed).
Run Python scripts as usual.
For other software (LM Studio, etc.), refer to Commands_and_Scripts.md and the screenshots for guidance.

📁 MCP_Server_Tools

Technologies: Model Context Protocol (MCP), Claude Desktop, GitHub Copilot, FastAPI, RSS
Goal: Build and inspect MCP servers for different scenarios:

Scenario 1: Basic calculator MCP
Scenario 2: FastAPI + MCP integration
Scenario 3: FreeCodeCamp RSS feed reader
Deployment examples

Also includes configuration for Claude Desktop and GitHub Copilot integration.

Key Images:

Claude Desktop Toy Dataset (1)	Claude Desktop Toy Dataset (2)

Copilot Calculator	Copilot Feed Search

MCP Inspect Scenario 1	MCP Inspect Scenario 2	MCP Inspect Scenario 3

Instructions:

Claude Desktop MCP:
1. Install Claude Desktop, enable developer settings.
2. Edit the config file using the provided sample_claude_desktop_config (adjust paths to your conda env).
3. Reload MCP in Claude.
GitHub Copilot MCP:
1. Install dependencies and the MCP inspector globally: npm install -g @modelcontextprotocol/inspector.
2. Follow the per‑scenario run commands (e.g., npx @modelcontextprotocol/inspector python MainCode/Scenario1/file.py).
3. Configure .vscode/mcp.json using sample_mcp.json.
4. Open GitHub Copilot in VS Code, select the model, and allow the MCP server.

📁 Smolagent_and_LangFuse_Apps

Technologies: Smolagents, LangFuse, Code agents, Multi‑agent, Retrieval, Tool calling, Vision
Goal: Dive into lightweight agent frameworks with observability:

Code agents, multi‑agent notebooks, retrieval agents, tool calling, tools, vision agents.

Key Images:

Code Agent	Multi‑Agent

Retrieval Agent	Tool‑calling Agent

Tools	Vision Agent

Instructions:
Run the notebooks in Google Colab (recommended for free GPU). Each notebook is self‑contained.

🎓 Certifications & Courses

This repository incorporates materials and projects from the following renowned courses:

Course	Completion Evidence
LangChain Academy Intro-to-LangChain
LangChain Academy Intro-to-LangGraph
LangChain Academy Intro-to-LangSmith
Hugging Face Agents Course

Place your certificate image links above.

The Intro-to-LangChain, Intro-to-LangGraph, and Intro-to-LangSmith folders contain the official tutorial notebooks from LangChain Academy. The Smolagent_and_LangFuse_Apps and other agentic projects reflect concepts taught in the Hugging Face Agents Course.

🚦 Getting Started

To replicate any project:

Clone the repository

git clone https://github.com/KraTUZen/Agentic-AI-and-Generative-AI-Practice.git
cd Agentic-AI-and-Generative-AI-Practice

Choose a project folder (e.g., CrewAI_Apps).
Set up a virtual environment
- Conda: conda create -n agentic-and-generative python=3.10
- UV: uv venv
- Pip: python -m venv venv
Install dependencies
```
pip install -r requirements.txt
```
(For LangBase apps, use npm install.)
Add environment variables
Create a .env file in the project folder with required API keys (see instructions.txt or notebook cells).
Run the code
- Python scripts: python script.py
- Jupyter notebooks: jupyter notebook notebook.ipynb
- TypeScript: npx tsx file.ts
- Streamlit apps: streamlit run app.py
For MCP servers, follow the detailed instructions inside the MCP_Server_Tools folder and the main instructions.txt.

🤝 Contributing

Contributions are welcome! If you have improvements, new agentic or generative ai projects, or bug fixes:

Fork the repository.
Create a feature branch.
Commit your changes.
Open a pull request.

Please ensure any added code includes clear instructions and screenshots where applicable.

📄 License

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

⭐ If you find this repository useful, please consider giving it a star!

Exploring the frontier of autonomous AI agents, one project at a time.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

🤖 Agentic AI and Generative AI Practice

📘 Introduction

📑 Table of Contents

⚙️ Technical Stack

🏗️ Repository Structure