Nebius AI + Hono + Cloudflare Example

This is a project to provide an example of using Nebius AI with Hono.

Features

• Minimal & Lightweight: Built with Hono and React for a streamlined, easy-to-follow example.
• Multi-Model Support: Easily switch between Nebius AI models like Deepseek R1, Meta-Llama, and others.
• Real-Time Streaming: Displays streaming responses from the AI in real time.
• Input Validation: Utilizes Zod for validating incoming data.
• Dual Rendering: Supports both server-side and client-side rendering.

Requirements

• Node.js v18+
• Nebius Account (For Inference Models)
• Cloudflare Account (for deploying to Cloudflare Pages/Workers)

Installation

1. Clone the Repository:

   git clone https://github.com/Arindam200/nebius-hono.git
   cd nebius-hono

2. Install Dependencies:

   npm install

3. Set Up Environment Variables:

   This project uses environment variables to configure your Nebius API key. You can set the variable using one of the following methods:

   • .env File: Create a .env file in the root directory (or copy from .env.example) with the following content:

     VITE_NEBIUS_API_KEY="Your NEBIUS API KEY"

   • .dev.vars File: When developing for Cloudflare Workers or if you prefer to keep your secrets separate, you can alternatively define your environment variables inside a .dev.vars file. This file is automatically detected by the Hono environment adapter. For example:

     NEBIUS_API_KEY="Your NEBIUS API KEY"

Development

Start the development environment using Vite:

npm run dev

Then open your browser at http://localhost:5173/ to interact with the chat application.

Deployment

To build and deploy the project to Cloudflare Pages, follow these steps:

1. Build the Application:

   The following command builds both the client and server bundles:

   npm run build

2. Deploy to Cloudflare Pages:

   Use the Wrangler CLI (configured via wrangler.toml) to deploy:

   npm run deploy

💡 Note:

In a production environment, ensure that the VITE_NEBIUS_API_KEY (or NEBIUS_API_KEY if using a .dev.vars file) environment variable is set in your hosting platform's environment settings.

Project Structure

.
├── .env                      # Local environment variables file
├── .env.example              # Example environment variables file
├── .dev.vars                # Development environment variables for Cloudflare Workers
├── package.json              # Project manifest and scripts
├── README.md                 # Project documentation
├── wrangler.toml             # Cloudflare Pages/Workers configuration
├── vite.config.ts            # Vite configuration file
└── src
    ├── client.tsx            # React-based client entry point
    ├── index.tsx             # Hono server/API entry point
    ├── renderer.tsx          # React renderer for Hono pages
    └── styles.css            # Application and component-specific styles

Configuration

Environment Variables

• VITE_NEBIUS_API_KEY / NEBIUS_API_KEY:
  This key is essential for your application to interact with the Nebius AI API. Use the VITE_NEBIUS_API_KEY variable in the .env file for local development with Vite or use NEBIUS_API_KEY in .dev.vars if you’re working with Cloudflare Workers.

Switching AI Models

By default, the project uses a Nebius AI model defined in src/index.tsx. To switch between models (e.g., from Meta-Llama to Deepseek R1), update the model parameter in the AI API call:

const result = await client.chat.completions.create({
  temperature: 0.6,
  model: "meta-llama/Meta-Llama-3.1-70B-Instruct", // Or change to "deepseek/R1"
  messages: messages,
});

This allows you to quickly swap models based on your current requirements.

Contributing

Contributions are welcome! If you have suggestions, improvements, or bug fixes, please open an issue or submit a pull request. Ensure that any contributions adhere to the project's style and testing guidelines.

License

This project is licensed under the MIT License. See the LICENSE file for the full license text.

Support

If you encounter issues or have any questions, please open an issue in the GitHub repository or contact the maintainers directly.