How I Prompted My Way to Protected Pages: AI as Scribe, Critic, and Tutor

Recently, I needed to add some protected routes to my Expo app. I have watched some good videos about the implementation. However, instead of the traditional "watch, pause, and type" method, I used an AI-augmented workflow that took me from a YouTube URL to running code in minutes.

Here is the 6-step process I used to bridge the gap between video tutorials and a stable implementation.

1. Source the Expert Knowledge

I started with a high-quality video tutorial on Expo Protected Routes. I’ve watched it before, so I knew the logic was sound, but I didn't want to manually scrub through the timeline to find the snippets I needed.

2. Extraction via Gemini

I pasted the YouTube URL into Gemini and asked it to:

• Summarize the core logic of the video.


• Extract a code sample based on the video’s implementation.

Reminder: I made sure to Like the video immediately. Since the creator’s content provided the "brain" for the AI's summary, it’s only fair to ensure they get the credit and engagement they deserve!

3. The "Second Opinion" Audit in Cursor

I took the code Gemini generated and moved it into Cursor. Instead of blindly pasting it into my project, I asked the AI:

"Is this code sample optimal for a sandboxed app environment? Are there any stability concerns?"

4. Architectural Pivot

This was the most valuable step. Cursor analyzed the code and suggested a more "stable" approach. While the video focused on a complex Protected Route layout, Cursor recommended using the Redirect component from expo-router.

5. Rapid Prototyping

I prompted Cursor to generate the final version using the Redirect pattern. It handled all the boilerplate—checking the global auth state and ensuring the user is redirected to /login if they aren't authenticated.

6. Active Learning & Verification

Once the code was running, I didn't just move on. I spent time:

• Running the app to test edge cases.


• Asking the AI to explain specific lines I didn't recognize.


• Refining the logic to fit my exact folder structure.

Why This Matters

I am moving away from "copy-pasting" and into "AI-orchestration." By using Gemini as a Researcher and Cursor as a Lead Architect, I was able to:

1. Save time on transcription.


2. Get a critique of the tutorial logic before it hit my codebase.


3. Choose the most stable path for my specific project needs.


4. Keep the momentum