☕ Starbucks Assistant — AI Chatbot Powered by LLaMA 3 & LangChain

Ask anything about Starbucks — Get intelligent, summarized, and relevant answers using LLaMA 3 (via Ollama), LangChain, and Retrieval-Augmented Generation (RAG) 🔍☕,also find the nearest Starbucks Works fully offline. Runs in your browser with a friendly coffee-themed UI.

🌟 Features

• 🤖 Ask Anything About Starbucks – From drinks and rewards to sustainability, store policies, and fun facts.
• 🧠 Powered by LLaMA3 (via Ollama) – Runs locally using Meta's LLaMA3-8B model with fast response and full control.
• 🔍 RAG-based Chatbot – Uses Retrieval Augmented Generation for accurate, context-rich responses.
• 🗺️ Real-Time “Starbucks Near Me” – Uses Google Maps API to show nearby Starbucks locations.
• 💬 Engaging Coffee-Themed UI – Chat UI styled with Starbucks color palette, emojis, and rounded messages.
• ☕ “Brewing” Typing Animation – Visual feedback while generating the answer.
• 📝 4-5 Sentence Answers – Optimized for clarity and brevity.
• ✅ Local & Private – Works entirely on your machine; no cloud APIs needed.

📦 Technologies Used

Tool / Library	Purpose
🦙 llama3:instruct	LLM from Meta, fine-tuned for instruction-based Q&A
🧠 Ollama	Framework to run LLMs locally with 1-line setup
🔗 LangChain	Chaining inputs, context and model logic
📚 LangChain-Chroma	Vector store to store & retrieve relevant reviews (retriever)
🌐 Flask	Lightweight Python web framework for hosting the chat UI
📊 Pandas	Load and manipulate the CSV review dataset
🧠 RAG	Retrieval Augmented Generation – smartly fetches relevant context
🎨 Tailwind CSS	Utility-first styling to match the coffee/Starbucks aesthetic
📁 Custom Dataset	Starbucks customer reviews (CSV file) used as knowledge base
🗺️ Google Maps API	To fetch real-time Starbucks nearby
🔐 python-dotenv	Load API key from .env

🧠 How It Works

1. You type a question.
2. The backend retrieves relevant Starbucks reviews using vector similarity from a Chroma vector store.
3. The question + top reviews are passed to the llama3:instruct model through LangChain.
4. The response is limited to 4-5 concise, informative sentences with some ☕ emojis for engagement.
5. The UI displays the result with a coffee-themed animation and style.

📁 Project Structure

Starbucks_Assistant/ │ ├── starbucks_reviews_data.csv # Dataset of Starbucks reviews (used in RAG) │ ├── app.py # Flask app serving the chatbot interface ├── main.py # CLI-based question tester for debugging ├── vector.py # Vector retriever using LangChain + Chroma │ ├── pycache │ ├── chroma_langchain_db │ └── chroma.sqlite3 │ ├── .env ├── venv/ │ ├── static/ │ ├── Starbucks-logo.png # Circular Starbucks logo │ ├── templates/ │ └── index.html # Chat UI page │ ├── Screenshots(Using UI)/ ├── Screenshots(without UI only , on local system)/ │ ├── requirements.txt # Python dependencies └── README.md # This file

🚀 How to Run

1. Clone the repository

git clone https://github.com/your-username/starbucks-assistant.git
cd starbucks-assistant

2. Install dependencies in a virtual environment

python -m venv venv
venv\Scripts\activate # On Windows
source venv/bin/activate # On mac/linux 
pip install -r requirements.txt

3. Download the model (LLaMA 3)

Ensure Ollama is installed and running, then:

ollama run llama3:instruct

4. Add your data file

Place your CSV file (starbucks_reviews_data.csv) in the vector.py path.

5. Add Google Maps API key

Create a .env file in the root folder with:

GOOGLE_MAPS_API_KEY=your_actual_api_key_here

6. Run the Flask App

python app.py

Visit: http://localhost:5000 in your browser.

💡 Sample Questions to Try

🥤 What are the most loved Starbucks drinks? 🍪 Do Starbucks stores sell food items too? 💰 What is the price range of Starbucks coffee? 🌿 Is Starbucks sustainable? 🎁 What are the Starbucks rewards? 🏪 What’s the vibe like in Starbucks stores? 📍 Starbucks near me

License

This repository is licensed under the All Rights Reserved terms.
You are free to view the code, but reproduction, modification, or reuse is not permitted without permission.

☕ Credits Built by Sai Jagdale ✨ Inspired by the joy of coffee, open-source LLMs, and LangChain innovation.