📝 AI-Powered Grammar & Spell Checker

📖 Check out my blog: Click Here

🚀 Try It Out

🔗 Live Demo: Click Here

🚀 Overview

This project is an AI-powered grammar and spell checker that utilizes Groq's LLM (Large Language Model) to correct spelling, grammar, and context errors in both English and Hindi. The application is built using Streamlit for an interactive user experience.

---

✨ Features

• ✅ Supports English and Hindi
• 🔍 AI-driven Grammar & Spell Correction
• 🎨 User-friendly UI with a dark theme
• ⚡ Instant Results with a single click
• 🏠 Built with Streamlit for easy deployment

---

🌟 Why Llama3-8B-8192?

This project uses Groq's Llama3-8B-8192 model, Meta's advanced large language model optimized for fast inference. It was chosen because:

• ✅ High Accuracy: It effectively detects and corrects grammatical errors while preserving context and meaning.
• ⚡ Lightning Fast: Groq's hardware acceleration provides ultra-fast response times, making it ideal for real-time applications.
• 💡 Superior Context Understanding: Llama3-8B excels at maintaining sentence coherence, nuance, and natural language flow.
• 📚 Efficient Processing: Can handle complex and lengthy text inputs with remarkable efficiency and accuracy.
• 🎯 Multilingual Support: Excellent performance in both English and Hindi language correction tasks.

---

🌍 How This Project Works

1. The user enters text that needs correction.
2. The app sends the text to Groq's LLM API, which processes the input using the advanced Llama3-8B-8192 model.
3. The LLM analyzes the text for grammar, spelling, and contextual errors.
4. The corrected text is then returned and displayed to the user.

🧠 How LLM Works in This Project

• Text Input: User provides raw text in English or Hindi.
• System Prompt: The LLM is instructed to act as a grammar and spell checker.
• Processing: The model detects errors and improves sentence structure while keeping the original meaning intact.
• Output: A refined and corrected version of the input is returned.

---

📝 System Prompt Used

The model is instructed with a carefully designed system prompt to ensure accurate and context-aware corrections. The prompts used are:

English Prompt

You are an advanced English grammar and spell checker. Correct the text while maintaining its meaning.

Hindi Prompt

आप एक अत्याधुनिक हिंदी व्याकरण और वर्तनी सुधारक हैं। 
कृपया दिए गए हिंदी पाठ को सही करें लेकिन उसका अनुवाद न करें। 
केवल व्याकरण, वाक्य संरचना और वर्तनी की त्रुटियों को ठीक करें, अर्थ को वैसा ही रखें।

🔄 Why These Prompts Are Used

• ✅ Ensures Precision: The prompts explicitly ask the LLM to focus on correcting grammar and spelling while maintaining context.
• 🌐 Supports Multiple Languages: Language-specific prompts allow optimal performance for English and Hindi.
• 💠 Preserves Meaning: The instruction ensures that the original intent of the text remains intact.
• 🎯 Optimized for LLM Processing: It minimizes ambiguity, making the model's responses more predictable and reliable.
• 🚫 Prevents Translation: Specifically instructs Hindi corrections without translation to English.

---

🛠️ Installation

🔹 Prerequisites

Make sure you have Python 3.8+ installed on your system.

🔹 Setup

# Clone this repository
git clone https://github.com/SimranShaikh20/grammar-checker.git
cd grammar-checker

# Install dependencies
pip install -r requirements.txt

# Run the application
streamlit run app.py

---

🌌 API Integration

This project uses Groq LLM API for text correction.

🔹 API Endpoint

POST https://api.groq.com/openai/v1/chat/completions

🔹 Example API Request

import requests

API_KEY = "your_groq_api_key"
text_to_correct = "Thiss is an exemple of baad grammr."
language = "English"

url = "https://api.groq.com/openai/v1/chat/completions"
headers = {"Authorization": f"Bearer {API_KEY}", "Content-Type": "application/json"}

payload = {
    "model": "llama3-8b-8192",
    "messages": [
        {"role": "system", "content": "You are an advanced English grammar and spell checker. Correct the text while maintaining its meaning."},
        {"role": "user", "content": f"Correct this English text: {text_to_correct}"}
    ],
    "temperature": 0.2,
    "max_tokens": 500
}

response = requests.post(url, headers=headers, json=payload)
print(response.json()["choices"][0]["message"]["content"])

---

🎨 UI Customization

This project includes custom CSS styling for a modern UI:

• 🌑 Dark mode background
• 🌟 Multicolored gradient buttons
• 🌍 Styled text inputs with improved readability
• 🛠️ Custom sidebar with a sleek layout

---

🔮 Future Enhancements

• 🚃 Voice Input Support
• 📚 Multi-language Support (Spanish, French, etc.)
• 🖊 Auto-correction while typing
• 🌐 Deploy as a Web App

---

🔄 Workflow

sequenceDiagram
    participant User
    participant Streamlit
    participant GroqAPI
    participant Llama3LLM
    
    User->>Streamlit: Enters text + selects language
    Streamlit->>GroqAPI: Sends correction request
    GroqAPI->>Llama3LLM: Processes with system prompt
    Llama3LLM->>GroqAPI: Returns corrected text
    GroqAPI->>Streamlit: Receives response
    Streamlit->>User: Displays corrected text

Loading

🐟 License

This project is licensed under the MIT License.

---

🎯 Author

Simran Shaikh

🚀 Made with ❤️ by Simran