My First Hands-On Experience with AWS Kiro
Recently, I explored Kiro, AWS's agentic coding environment, and used it to build a simple Feedback Form with Name, Email, and Message fields. Let's see what made me interested in AWS Kiro.
Introduction:
As developers, we use many AI tools to write code faster. But I think coding is only one part of building a project. What happens if we understand the complete idea before building a project like understanding the requirements, planning the solution, testing it, and keeping the project organized?
That's what made me interested in AWS Kiro.
For my first hands-on project, I decided to build a simple Feedback Form with Name, Email, and Message fields.
Note: A simple feedback form doesn't really need this much process on its own - I just used it as an example for demo purposes, to explain how Kiro's workflow works in practice.
What is AWS Kiro?
Kiro is an AWS-provided, AI-powered development environment IDE that helps turn an idea into a structured development process.
The basic workflow:
Idea -> Requirements -> Design -> Tasks -> Implementation -> Testing
This is also called spec-driven development in AWS.
Instead of directly asking AI to write code, I could first understand what I actually needed to build and how I planned to build it.
- Download Link: Downloads - Kiro
- Sign in with your AWS Builder ID afterward
- Click on your profile, and you'll see 50 credits per month on the Free plan.
- That means monthly you can use only 50 credits, which is more than enough.
Step 1: Requirements
- This is how the AWS Kiro IDE looks when I started with a simple prompt
Create a feedback form with Name, Email, and Message, validate the inputs, and show a thank you message after submission
- Kiro created a
requirements.mdfile.
It included requirements such as:
- Name, Email, and Message fields
- Input validation
- Error messages
- Character limits
- Loading state
- Thank-you screen
- Submission error handling, and more
This helped me clearly understand what the application should do before writing code
Step 2: Design phase
- Make sure the build-in option is selected in the spec and choose the model you prefer. I'm comfortable with Claude Sonnet 4.5, so I used that.
- Next, Kiro created a design.md file
- The requirements.md file describes what we should build, whereas the design.md file describes how we should build it.
It explained the plan, like:
- Project structure
- Components
- Validation logic
- Form handling
- Error handling
- Testing approach, etc.
Step 3: Tasks
- Kiro then created a tasks.md file and divided the work into smaller tasks.
- You have to click " Start Task " for each step and also allow the libraries and commands to run similar to what you'd see in Claude or Antigravity.
In the tasks.md file, you'll see small tasks, for example:
- Set up the project
- Create validation functions
- Validate Name, Email, and Message
- Add tests
- Check character limits
- Handle invalid inputs
This is because instead of looking at one big problem, I could work with small, manageable tasks.
Step 4: Implementation
- After reviewing the tasks, I used "Run All Tasks"
- Kiro started working through the tasks and showed the progress.
This was the part I found most interesting, because I could see the process:
Plan -> Build -> Test -> Improve
What Problem Does Kiro Solve?
One common problem I noticed while using AI coding or vibe-coding tools is that we can sometimes focus too much on code
like just saying "Give me the code."
But for big or larger projects, we also need proper planning and structure.
That's where AWS Kiro helps, it gives a structured approach that keeps the requirements, design, and tasks together with the project.
That makes the project easier to understand and maintain
Why Kiro is Different from Claude and GitHub Copilot
I wouldn't say one tool is better than another, they're useful in different ways.
| Tool | My experience |
|---|---|
| Claude | Reasoning, explanations and coding |
| GitHub Copilot | Coding assistance while developing |
| Kiro | Planning and building features through a structured workflow |
The biggest difference I noticed with Kiro was the spec-driven workflow.
What I Learned from this
- This small project taught me that:
AI-assisted development is not only about writing code faster.
It is also about:
Understand -> Plan -> Build -> Test -> Improve
- AI can help us with the repetitive work, but we still need to understand the workflow and review what's being built.
Final Thoughts
I started this project with just one simple idea.
Create a feedback form
But Kiro helped me turn that idea into:
Idea -> Requirements -> Design -> Tasks -> Implementation -> Testing
That was my biggest takeaway from this experience.
I am not trying to say AI should replace developers. Instead, I'm learning how to become a better developer by working with AI.
For more information, you can also read the AWS documentation here: Kiro: Move Beyond AI Coding to Agentic Engineering
Conclusion:
I'm continuously learning about new AI updates on AWS through hands-on projects, learning, and sharing my experiences.
My main goal is to keep following:
Learn -> Build -> Document -> Share -> Help to Others
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article