Building Web Apps with Bolt

Chapter 3. Be Clear, Be Specific: Prompt Engineering and Iterative Development

Now that you’ve seen an app come to life from a single prompt, it’s time to dive deep into the art and science of writing prompts. This chapter will show you how to get the best results and how to manage the output the AI gives you. Think of this as learning to communicate with your apprentice—they’re extremely smart, fast, and knowledgeable, but they need clear instructions and a bit of guidance.

I’ll cover techniques for crafting good prompts, refining your app’s behavior and design through natural language, and handling the AI’s responses (including when it doesn’t do exactly what you want), and I’ll show you some best practices for a smooth prompt-first workflow. By the end of this chapter, you should feel confident in explaining what you want (even if you don’t know how to code it yourself) and directing Bolt to build it.

Managing Expectations: AI’s Limitations and Strengths

As you continue to work with Bolt, it’s important to maintain realistic expectations and leverage the AI’s strengths. Bolt is incredibly powerful at generating code quickly and following instructions, but it’s not infallible or all-knowing. It can’t instantly create a perfect, production-ready app without your input and guidance. Sometimes it will make an assumption that you didn’t intend, and it won’t know what’s on your mind unless you tell it explicitly.

One limitation to be aware of is the AI’s memory, or context window ...