Ollama Coder: AI-Powered Code Generation

Ollama Coder is a web application that generates React components from natural language prompts. It leverages the power of the Ollama API to translate your descriptions into functional, styled TypeScript code.

Features:

• Natural Language Input: Describe the React component you need in plain English.
• AI-Powered Code Generation: Ollama's powerful models translate your prompt into working code.
• TypeScript Support: Generated code is written in TypeScript for type safety and maintainability.
• Tailwind CSS Styling: Components are automatically styled using Tailwind CSS classes for a consistent and professional look.
• Interactive Components: The generated code includes state management where necessary to create fully functional components.
• Model Selection: Choose between different Ollama models (qwen2.5-coder:14b, phi3:14b are recommended) for varied code generation capabilities. (See instructions below on how to add more models).
• Streaming Response (Implicit): The application fetches and displays the generated code progressively as it's received from the Ollama API.

Technology Stack:

• Frontend: Next.js, React, TypeScript, Tailwind CSS, Radix UI, Framer Motion
• Backend: Next.js API routes
• AI Model: Any LLM model support by Ollama. By default uses qwen2.5-coder:14b and phi3:14b
• Code Display: Custom CodeViewer component

Getting Started (For Developers):

1. Clone the repository: git clone https://github.com/luongnv89/ollama-coder
2. Install dependencies: npm install
3. Set up Ollama API: The API endpoint (ollamaApiUrl) is not included in this repository and needs to be configured within the route.ts file. Refer to the Ollama documentation for setting up your API key and endpoint.
4. Run the development server: npm run dev

Adding More Models:

To add more models to the available options, modify the models array in page.tsx. Ensure that the value property accurately reflects the name of the model as it's recognized by the Ollama API.

Important Considerations:

• This project is a personal project and is not officially affiliated with Ollama.
• The quality of the generated code depends on the clarity and precision of your prompt.
• Error handling is implemented to catch issues with API requests, but unexpected behavior may still occur.

Contributing:

Contributions are welcome! Please open an issue or submit a pull request.

License:

MIT License (See LICENSE file)

This project is inspired by geminicoder.