Hey Everyone, welcome to our new Build with AI Series. Let’s explored most efficient way to use Cursor AI with minimal mistakes. If you're building your Artificial Intelligence product then must follow this guide for most efficient time-saving results.

1/ Plan Before Coding

Before you start coding, always make a clear plan. Use Cursor’s V0 to create mockups and get an idea of your project before writing any code.

2/ Leverage ChatGPT for Project Planning

Ask ChatGPT to help you with things like creating your project requirements (PRD), database design, color palettes, and structural advice. Save all this information in .md files inside Cursor to keep everything organized. This way, you’ll have a clear roadmap for your project.

3/ Use cursor.directory for Better Prompts

Cursor’s directory offers tech-specific prompts. You can copy and edit these to fit your project, then save them in a .cursor_rules file. This makes the AI’s responses more accurate, especially when customized to your tech stack.

4/ Tag Relevant Docs

Sync official documentation like Next.js or Supabase into Cursor. This helps ensure you’re always getting the latest, most accurate answers, instead of relying on old or outdated solutions.

5/ Save Working Code in .md Files

Whenever Cursor provides code that works for your project, save it in a .md file. This way, you can easily go back to it later, and Cursor will give faster, more accurate suggestions based on that code.

6/ Use AI to Explain Code

You can also use Cursor to learn. Ask it to explain your code in simple terms, like you're a beginner. This helps you spot patterns and understand how your code works better over time.

7/ Start with Code Templates

Don’t spend time creating everything from scratch. Use ready-made templates for common things like authentication, databases, and payments. You can find these templates on sites like vercel.com. This saves time and ensures you're starting with reliable code.

Wrapping Up!

Give Cursor as much context and information as possible, and you’ll notice fewer mistakes and better efficiency.

Next Read: Tips for Building MVP with CursorAI

FAQs

1. What Large Language Model (LLM) do you use?

I use Sonnet 3.5 for coding tasks. For project planning and other non-coding tasks, I prefer using GPT-4o.

2. My projects are getting bigger, and Cursor seems to forget previous specifications and starts hallucinating. I’ve tried using a pre-prompt to remind it of everything—specs, goals, code structure, etc. Is this useful?

Yes, Cursor can sometimes forget previous inputs. To improve this, I recommend creating different .md files for each aspect of your project, like PRD.md, db-design.md, and design-principles.md. Each time you're working on something new, briefly explain what you’ve done so far and refer Cursor to these files. It significantly improves the tool's accuracy.