ToneSlider Backend

The backend server for ToneSlider, a tone adjustment tool that allows users to rewrite text in varying levels of formality using a slider. This backend connects with the Mistral AI API to process text tone, uses Upstash Redis for response caching, and is designed to be fast and efficient.

🌐 Live API

The backend is deployed at:
https://toneslider.up.railway.app

🛠️ Tech Stack

• Node.js with Express.js – REST API server
• Upstash Redis – Caching layer for performance
• Mistral AI API – Language model for tone transformation
• CORS – For secure cross-origin requests
• dotenv – Environment variable management

📦 Features

• Accepts text and tone level, returns rewritten text in the desired tone
• Uses caching to reduce API calls and improve response time
• Error handling and graceful fallbacks
• Environment-variable-driven setup for secure credentials

🚀 Getting Started

1. Clone the repository

git clone https://github.com/yourusername/toneslider-backend.git
cd toneslider-backend

2. Install dependencies

npm install

3. Create a .env file

PORT=3000
MISTRAL_API_KEY=your_mistral_api_key
REDIS_URL=your_upstash_redis_url

4. Start the server

npm start

📤 API Endpoint

POST /adjust-tone

Adjusts the tone of a given piece of text based on the tone level.

Request Body:

{
  "text": "Your original text here.",
  "toneLevel": 70
}

Response

{
  "content": "Your text rewritten in a casual tone."
}

---

🧠 Caching with Upstash Redis

To improve performance and reduce redundant API calls to Mistral, this backend uses Upstash Redis for caching.

🔧 How it Works

• Each request to adjust tone is cached using a unique key format:
• If a cached result exists, it is returned instantly without making an API call.
• Cached data has a TTL (Time To Live) of 3600 seconds (1 hour).

✅ Benefits

• Reduces latency for repeated requests
• Lowers Mistral API usage and costs
• Provides a smoother user experience

---

⚠️ Error Handling & Edge Cases

This backend handles errors gracefully to ensure users aren't left stuck:

• API Failures: If the Mistral API fails or returns an error, the server returns the original input text with an error message.
• Redis Failures: If Redis is unreachable or fails to connect, the system continues to function but skips caching.
• Invalid Input: Requests without valid text or toneLevel are ignored with appropriate client-side validation.
• Frontend Feedback: Using sonner toasts, users are alerted when:
• Tone adjustment fails
• An undo or redo operation is performed
• Text is reset to its original state

---

🛠️ Contribution

We welcome contributions to improve ToneSlider! If you'd like to help:

1. 🍴 Fork the repository
2. 🛠️ Create a new branch (git checkout -b feature-name)
3. 💾 Make your changes and commit (git commit -m "Add some feature")
4. 🚀 Push to your branch (git push origin feature-name)
5. 🔁 Open a Pull Request

Guidelines

• Write clean, readable code with comments if needed.
• Ensure all tests pass before submitting.
• Keep commits atomic and descriptive.
• For new features or bug fixes, please open an issue first to discuss the idea.

📦 Frontend Setup:
To view and interact with the full ToneSlider application, please refer to the https://github.com/isid555/ToneSlider and follow its setup instructions.

📫 For major changes or feature discussions, feel free to reach out or open an issue. Let’s build something amazing together!

Thank you