🎃 FixIt AI — How I Built a Full AI Repair Assistant in One Week Using Kiro

Source: Dev.to

👻 Inspiration — When a Simple Idea Meets Real Need

I didn’t build FixIt AI because it was “cool.” I built it because I needed it — and people around me needed it too.

Where I live, many device issues (slow laptop, flickering screen, draining battery) don’t actually require a technician. But the knowledge gap means people pay every time. As a 17‑year‑old developer passionate about solving real problems, I wanted to build a tool that:

• empowers users with instant diagnostics
• reduces unnecessary repair costs
• helps young people and students avoid “tech emergencies”
• uses AI to make hardware more accessible

Kiroween gave me the creative angle: a funny, spooky ghost assistant that “haunts away” your device problems. FixIt AI was born.

🛠️ What FixIt AI Does

FixIt AI analyzes symptoms like:

• “My laptop is overheating”
• “My phone won’t charge”
• “My microwave is sparking”

It then generates:

• structured diagnosis
• severity rating
• estimated cost
• whether a technician is needed
• beginner, intermediate, and advanced repair steps
• safety warnings
• device lifespan notes
• probability of success
• spooky flavor text
• a glowing ghost animation while processing

Tech stack

• Frontend: Netlify (HTML, CSS, JS)
• Backend: Node.js + Express on Render
• AI logic: Kiro‑assisted diagnosis engine

⚙️ How I Used Kiro to Build FixIt AI

🟣 1. Vibe Coding — My Main Workflow

I built FixIt AI conversationally, using Kiro like a pair‑programming teammate. Kiro helped me:

• rewrite complex frontend functions (e.g., showResult())
• design the multi‑tier repair instructions structure
• write spooky‑themed UI text and animation concepts
• debug deployment errors and Express routing issues
• build the diagnosis engine logic
• optimize JSON responses
• refine prompts for better reasoning

🟣 2. Spec‑Driven Development

Before coding the engine, I defined:

• required JSON structure
• expected fields (severity, probability, steps)
• multi‑layer outputs
• UI expectations
• error‑handling behavior

I handed this spec to Kiro, which generated consistent logic, speeding development and preventing contradictory outputs.

🟣 3. Steering

I used steering prompts such as:

• “Rewrite this completely to be more professional.”
• “Improve the output, make it spookier but still readable.”
• “Make this judge‑friendly.”
• “Simplify; now restructure it; now enhance it.”

Each round aligned Kiro more closely with my needs.

🟣 4. Rapid Prototyping & Debugging

Kiro helped me solve:

• CORS errors
• Render’s Node version failure
• Express incorrect dependency path
• Frontend/backend connection mismatch
• CSS animation placement issues

🟣 5. Creativity + User Experience

Kiro also shaped:

• the ghost loading animation
• the glowing UI
• the spooky diagnostic responses
• problem‑to‑solution storytelling
• architecture diagrams
• day‑by‑day plan

FixIt AI ended up looking more polished than anything I expected to finish within a hackathon.

🚀 What I Learned

• How to combine AI reasoning with traditional backend logic
• How to structure complex JSON outputs
• How to deploy full‑stack apps (Netlify + Render)
• How to design a delightful UX under time pressure
• How to work collaboratively with an AI assistant
• How powerful a spec‑driven approach can be
• How much Kiro accelerates development compared to coding alone

😅 Challenges I Faced

• Debugging Express on Render
• Designing a spooky theme that still feels usable
• Making outputs structured instead of random
• Rewriting results UI multiple times
• Time pressure — I had only one day to finish everything

Kiro helped me push through every block.

🎯 Final Thoughts

FixIt AI isn’t just a Halloween‑themed app. It solves a real problem — one that affects real people every day. Kiroween taught me that even with limited resources, a solo developer can build something useful, beautiful, and impactful when paired with the right AI tools.

This is just the beginning. I plan to expand FixIt AI into something even bigger — and bring affordable diagnostics to anyone with a device.

🔗 Try It

• Frontend:
• Backend:
• Code:

---