🦜 LangChain Playground for LLMs

The most complete, hands-on LangChain learning repository on GitHub

Build, learn, and ship production-ready LLM applications — from your first prompt to full-scale RAG pipelines and autonomous AI agents.

⭐ If this repo helps you, please give it a star — it helps others discover it too! ⭐

---

🚀 Why This Repository?

Most LangChain resources are scattered across tutorials, notebooks, and outdated blog posts. This repository brings everything into one place — structured, runnable, and ready for production.

🎯	What You Get
📚	Structured learning path — progress from zero to LLM expert in a logical sequence
💻	100% runnable code — every concept has a working example you can run immediately
🔍	RAG pipelines — full retrieval-augmented generation with Chroma, FAISS, and more
🤖	AI Agents — ReAct-pattern agents with custom tools and LangGraph workflows
🏠	Local LLM support — run everything offline with Ollama (zero API costs)
🌐	Multi-provider — OpenAI, Google Gemini, Hugging Face, and Ollama side-by-side
📝	Rich notes & docs — detailed markdown notes for every topic covered
🔧	Production patterns — LCEL, tool calling, structured output, and memory

---

Table of Contents

• Why This Repository?
• Learning Roadmap
• Repository Structure
• Notes Index
• Getting Started
• Local LLM Support (Ollama)
• Tech Stack
• Security
• Future Enhancements
• Topics & Keywords
• Contributing
• Connect

---

Learning Roadmap

The content is organized to be followed in order:

#	Topic	Description
1	Chat Models	LLM initialization with Ollama, Gemini, and OpenAI (chatmodel/)
2	Messages	SystemMessage, HumanMessage, AIMessage, and chat history (Messages/)
3	Prompt Engineering	Prompt templates, dynamic generation, and a Streamlit prompt UI (Prompts/)
4	Structured Output	Enforcing JSON schema, Pydantic, and TypedDict outputs (Structured Output/)
5	Output Parsers	JSON, Pydantic, string, and structured output parsers (Output_Parsers/)
6	Chains	Sequential, parallel, and conditional chains (Chain/)
7	Runnables (LCEL)	LangChain Expression Language — sequences, parallelism, branching, lambdas, passthrough (Runnables/)
8	Embeddings	Document similarity using embedding models (EmbeddingModel/)
9	RAG Components	Document loaders, text splitters, vector stores (Chroma), and retrievers (Rag Components/)
10	Tools & Agents	Custom tool creation, tool calling, built-in tools, and agent workflows (Tools/)
11	Mini Projects	Applied projects — AI agents, currency conversion, agent templates (Mini_projects/)

---

Repository Structure

langchain-playground--for-llms-/
├── Chain/                    # Sequential, parallel, and conditional chain examples
├── chatmodel/                # LLM initialization (Ollama, Gemini)
├── chroma_db/                # Persisted Chroma vector database
├── EmbeddingModel/           # Document similarity with embeddings
├── Messages/                 # Chat message types and prompt templates
├── Mini_projects/            # End-to-end LLM applications
├── Output_Parsers/           # JSON, Pydantic, string, and structured parsers
├── Prompts/                  # Prompt templates and Streamlit UI
├── Rag Components/           # RAG pipeline modules
│   ├── Document Loaders/     #   PDF, CSV, text, web, and directory loaders
│   ├── Retrievers/           #   MultiQuery, contextual compression, MMR
│   ├── Text Splitters/       #   Length, markdown, code, and semantic splitting
│   └── Vectors Stores/       #   Chroma vector database integration
├── Runnables/                # LCEL runnables (sequence, parallel, branch, lambda)
├── Sample Data/              # Reference PDFs for RAG examples
├── Structured Output/        # JSON schema, Pydantic, and TypedDict outputs
├── Tools/                    # Custom and built-in tool calling, agent toolkits
├── notes.md                  # Core LangChain concept notes (Chat Models, Prompts, Messages)
├── requirements.txt          # Python dependencies
├── test.py                   # Quick-start test script
├── sample.pdf                # Sample document for loader testing
├── sample_data.md            # Sample text data for RAG testing
├── Titanic-Dataset.csv       # Sample dataset for data-loading examples
└── README.MD

---

Notes Index

All concept notes are organized by topic. Click any link to jump directly to the notes.

🧠 Foundations

Topic	Notes File
Chat Models, Prompts & Messages	notes.md
Chat History	Messages/chat_history.md

---

⛓️ Chains

Topic	Notes File
Simple & Sequential Chains	Chain/chain.md
Parallel Chain	Chain/parallel_chain.md
Conditional Chain (IF/ELSE routing)	Chain/conditional_Chain.md

---

🔄 Runnables (LCEL)

Topic	Notes File
Runnables, LCEL, Primitive vs Task-Specific	Runnables/Runnables.md

---

📤 Output Parsers & Structured Output

Topic	Notes File
Output Parsers Overview & Ollama	Output_Parsers/notes1.md
String Output Parser	Output_Parsers/stringoutputParser.md
JSON Output Parser	Output_Parsers/jsonoutput_parser.md
Pydantic Output Parser	Output_Parsers/PydanticOutputParser.md
Structured Output Parser (legacy)	Output_Parsers/Structured Output Parser.md
TypedDict vs Pydantic vs JSON Schema	Structured Output/structured.md

---

📚 RAG Components

Topic	Notes File
Document Loaders (PDF, Web, CSV, Directory)	Rag Components/rag_notes.md
Text Splitters	Rag Components/text_splitter_langchain_notes.md
Vector Stores Overview	Rag Components/Vectors Stores/vector_store_notes.md
Chroma DB	Rag Components/Vectors Stores/chroma_db_notes.md
Chroma Internal Architecture	Rag Components/Vectors Stores/chroma_internal_architecture.md
FAISS vs Chroma vs Pinecone Comparison	Rag Components/Vectors Stores/vector_db_architecture_comparison.md
Retrievers Overview	Rag Components/Retrievers/retriver.md
MMR Retriever	Rag Components/Retrievers/mmr.md
MultiQuery Retriever	Rag Components/Retrievers/MultiQuery_Retriever_Notes.md
Contextual Compression Retriever	Rag Components/Retrievers/Contextual_Compression_Retriever_Notes.md

---

🔧 Tools & Agents

Topic	Notes File
Tools Overview	Tools/tools.md
Tool Binding	Tools/tool_binding.md
Tool Calling	Tools/tools_calling .md
Custom Tools	Tools/custom_tools.md
Toolkits	Tools/toolkit.md
AI Agents (ReAct Pattern)	Mini_projects/agent_notes.md

---

📄 Sample Data

File	Description
sample_data.md	AI overview text for RAG testing
sample.pdf	Sample PDF for document loader testing
Titanic-Dataset.csv	Sample CSV for data loading examples

---

Getting Started

Prerequisites

• Python 3.10 or higher
• Ollama (optional, for local LLM inference)

1. Clone the Repository

git clone https://github.com/himanshu231204/langchain-playground--for-llms-.git
cd langchain-playground--for-llms-

2. Create a Virtual Environment

python -m venv venv
source venv/bin/activate        # macOS / Linux
venv\Scripts\activate           # Windows

3. Install Dependencies

pip install -r requirements.txt

4. Configure Environment Variables

Create a .env file in the project root:

OPENAI_API_KEY=your_openai_key
GOOGLE_API_KEY=your_google_api_key
HUGGINGFACEHUB_API_TOKEN=your_hf_token

Note: The .env file is listed in .gitignore and should never be committed.

---

Local LLM Support (Ollama)

This repository supports local LLM inference via Ollama, allowing experimentation without API costs or internet access.

Setup

# Install Ollama from https://ollama.com, then:
ollama pull llama3
ollama run llama3

Usage with LangChain

from langchain_ollama import ChatOllama

llm = ChatOllama(model="llama3")
response = llm.invoke("Explain Retrieval Augmented Generation in simple words")
print(response.content)

Benefits of local inference:

• No API costs
• Full data privacy
• Offline development
• Faster iteration for prototyping

---

Security

• API keys are managed through environment variables (.env)
• .env is excluded from version control via .gitignore
• No secrets are hard-coded in source files

---

Tech Stack

Layer	Technologies
LLM Providers	OpenAI GPT-4, Google Gemini, Ollama (Llama 3, Mistral), Hugging Face
Frameworks	LangChain v1.0, LangGraph v1.0, LangChain Expression Language (LCEL)
Vector Stores	Chroma DB, FAISS
Embeddings	OpenAI Embeddings, HuggingFace Sentence Transformers
Output Validation	Pydantic v2, TypedDict, JSON Schema
UI	Streamlit
Data	Pandas, CSV, PDF, Markdown document loaders
Language	Python 3.10+

---

Future Enhancements

• Multi-agent systems with LangGraph
• Conversational memory integration
• Evaluation and monitoring pipelines
• Deployment-ready GenAI applications
• Streamlit-based interactive demos

---

Topics & Keywords

This repository covers the following areas of Generative AI and LLM development:

langchain llm generative-ai rag retrieval-augmented-generation langchain-python ai-agents openai google-gemini ollama vector-database chromadb pydantic lcel langchain-expression-language huggingface streamlit python machine-learning nlp large-language-models prompt-engineering ai deep-learning chatbot langsmith langgraph embeddings transformers tutorial

---

Contributing

Contributions are welcome. To get started:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/my-feature)
3. Commit your changes (git commit -m "Add my feature")
4. Push to the branch (git push origin feature/my-feature)
5. Open a pull request

---

Connect

• LinkedIn
• GitHub
• Twitter / X
• Buy Me a Coffee

---

🌟 Found this useful? Star the repo and share it with fellow AI enthusiasts!

Made with ❤️ by Himanshu — Happy Learning! 🚀