🦙 AI-chat-app-using-Groq-LLaMA-4-Scout

A sleek, fast AI chat interface built with Streamlit, powered by Groq's inference engine and Meta's LLaMA 4 Scout model. Ask anything and get blazing-fast responses with a clean dark UI.

---

✨ Features

• ⚡ Ultra-fast inference via Groq's hardware-optimized API
• 🧠 Multi-model support — LLaMA 4 Scout, LLaMA 4 Maverick, LLaMA 3, Mixtral
• 🎨 Custom dark UI — styled with CSS for a polished chat experience
• 🔧 Configurable settings — temperature, max tokens, and model selection via sidebar
• 💬 Context-aware conversations — full message history sent to the model each turn
• 🗑️ Clear chat — reset the conversation anytime

---

🛠️ Tech Stack

Layer	Tool
Frontend	Streamlit + Custom CSS
LLM Backend	Groq API
Model	meta-llama/llama-4-scout-17b-16e-instruct
Language	Python 3.8+

---

🚀 Getting Started

1. Clone the repo

git clone https://github.com/your-username/llama4-scout-chat.git
cd llama4-scout-chat

2. Install dependencies

pip install streamlit groq

3. Get your Groq API key

Sign up at console.groq.com and generate a free API key.

4. Run the app

streamlit run groq_chat_app.py

Then open http://localhost:8501 in your browser.

---

🔑 API Key Setup

You can provide your Groq API key in two ways:

Option A — Via the sidebar (recommended): Enter it directly in the app's sidebar under Groq API Key.

Option B — Via environment variable:

export GROQ_API_KEY="gsk_your_key_here"
streamlit run groq_chat_app.py

---

⚙️ Configuration Options

Setting	Range	Default	Description
Model	5 options	LLaMA 4 Scout	The LLM model to use
Temperature	0.0 – 2.0	1.0	Controls response creativity
Max Tokens	256 – 4096	1024	Maximum response length

---

📁 Project Structure

llama4-scout-chat/
│
├── groq_chat_app.py   # Main Streamlit app
└── README.md          # Project documentation

---

🤖 Supported Models

Model	Best For
meta-llama/llama-4-scout-17b-16e-instruct	Fast, general-purpose Q&A
meta-llama/llama-4-maverick-17b-128e-instruct	Long context tasks
llama3-70b-8192	High-quality detailed responses
llama3-8b-8192	Lightweight, quick replies
mixtral-8x7b-32768	Multilingual & long documents

---

📌 Notes

• The app displays only the latest question and answer on screen, keeping the UI clean — while still sending the full conversation history to the model for context.
• API keys are never stored permanently; they exist only in the current session.

---

Built with 🧡 using Groq + LLaMA 4 + Streamlit