I’m building a single web app that lets a user enter one prompt and instantly view how different large-language-model chatbots respond side-by-side. At a minimum the platform must connect to OpenAI’s GPT family and Google’s Gemini, with the architecture open enough for me to plug in additional APIs later. The core workflow is comparison: a prompt is sent concurrently to every connected model, the replies appear in parallel panels, and the user can scroll, copy, or highlight differences. To make the tool genuinely useful I also need these additional capabilities: • Save chat history per user so previous prompts and multi-turn conversations can be revisited. • Export comparison data in a convenient format (CSV, JSON or Markdown) for downstream analysis or sharing. • A customizable interface where users can reorder model panes, toggle dark/light themes, and adjust font size. You are free to choose the tech stack, but a modern JavaScript framework (React, Vue, or Svelte) on the front end and a lightweight Node/Python backend with websocket support feels right. Clean code, clear API separation, and straightforward instructions for adding new providers will be part of the acceptance criteria.