Muppala-Jasvanth-Varma/ai-video-generator-app — reverse-engineered prompt

Build me an Android app called PastPortals that lets someone type a simple history question like “What happened in 1947” or “Who was Nikola Tesla” and turns the answer into a short visual story.

The app should have a clean Kotlin and Jetpack Compose interface where users can enter a topic, see loading progress, then view the generated images and final short video. The backend should fetch reliable information from Wikipedia, summarize it into a clear historical narrative, create good prompts with Gemini, then send those prompts to an AI image or video generation service to produce the visuals. Keep the flow simple, query, research, prompt, generate, show result.

Please make the Node and Express backend and Android frontend work together, with sensible error handling if Wikipedia or the AI service fails. Use placeholders for API keys and clear setup notes so I can run it locally. Look up current docs online if you need to.