Fix AI App Image Rendering

My AI-based app builder already converts user prompts into clean UI layouts, but the images that should accompany those layouts are often wrong. Everything is stored and fetched from cloud storage right now, so the problem sits somewhere between the prompt-to-image mapping and the rendering layer. Here is what I need from you: • Dive into the existing repo, trace the current image-selection logic, and pinpoint why the wrong assets are being pulled. • Refactor or replace that logic so the simulator consistently shows images that match each prompt description. • Keep the solution storage-agnostic—whether the project ends up on AWS S3, Google Cloud Storage, or Azure Blob, the previews must remain accurate. • Optimise for speed and memory so the mobile UI simulator feels instant and lightweight. Acceptance criteria 1. Given a set of test prompts I provide, every preview must display the correct images on first load without manual refresh. 2. Switching cloud buckets or keys should require, at most, a single config change. 3. All new or modified code is clean, commented, and covered by lightweight tests. If that sounds straightforward to you, let’s get this fixed so users see exactly what they imagined.

Project ID: 40261492