This is a series of blog post from Tekai's personnel sharing their past working experience and how those experience can help Tekai's future clients.

Why Did I Choose This AI and Model?

While working with code, I realized that AI can be very helpful in various tasks, such as generating functions for specific purposes, improving the efficiency of existing code, providing ideas for handling specific issues in a code segment, and performing simple tasks. After experimenting with several AI-powered coding assistants (GitHub Copilot, Windsurf, etc.), I decided to use Cursor in combination with the Claude model.

Reasons for This Choice:

• CursorAI: This is an AI-powered code editor optimized for smart code suggestions. It saves time with strong autocomplete capabilities, a deep understanding of code context, and integration with an AI agent.

• Claude: Compared to GPT, Claude tends to respond in a more understandable way and avoids basic logical errors. It is suitable for handling simple tasks such as writing small scripts, modifying configuration files, or quickly generating code templates.

What Tasks Is This AI Suitable For? Who Should Use It?

Cursor + Claude is most useful in the following scenarios:

Developers looking to speed up their work: If you frequently write repetitive code or need help with syntax, AI can save you time.

Developers learning a new language or framework: Cursor and Claude handle simple tasks well, making them useful for beginners who need quick examples when learning a new language or framework. (I learned Flutter and Swift using Cursor without having to read the full documentation for these frameworks.)

Usage Costs

Cursor offers a free version with basic features, but full AI support requires a paid plan ($20/month).

Claude has free and Pro versions (monthly subscription). Claude is already integrated into the Cursor Editor, so if you have a Cursor subscription, no additional cost is needed.

Effective Usage in Different Contexts

Here are some ways I commonly use this AI in my work:

• Quickly Generating Boilerplate Code

When I need to create an API route, process a JSON file, or migrate data.

Example Prompt: "Write a JavaScript script to modify file ABC by replacing all occurrences of class="cls-1" with fill="#ed1c24"."

• Optimizing and Refactoring Code

When I want to make my code cleaner and more maintainable.

Example Prompt: "Optimize this code to make it shorter and easier to read."

• Generating Unit Tests Quickly

When I need to test functionality without writing tests manually.

Example Prompt: "Write a Jest test for this function."

• Handling Simple Tasks

Example Prompt: "Show a modal displaying user information when this button is clicked."

Example Prompt: "Instead of using Google Dynamic Links, migrate the code in this repo to use AppsFlyer OneLink."

Conclusion

Although it’s just one tool in a developer’s toolkit, using Cursor + Claude has helped me recognize the importance of documentation, team communication, and collaboration. For successful feature delivery, these factors must be closely monitored and continuously refined.

I hope these insights will be valuable for teams looking to improve product quality! 🌟