Vision – AI Chat Assistant UI

Vision is a clean, powerful AI chatbot interface designed to give users a beautifully minimal way to interact with AI. Whether you’re building the next ChatGPT, Perplexity, or something weird and experimental, Vision gives you the perfect starting point, complete with full functionality, chat memory, actions, and a futuristic vibe that makes people want to use it.

Build a modern AI chat assistant called Vision. The layout should be focused on clarity, minimalism, and performance. It should support real-time conversation, multiple assistant “modes” (like helper, researcher, therapist), and conversation history. Use dark mode as default, with a soft futuristic UI. The assistant should feel like a calm, competent partner, think less clippy, more Star Trek. Include animations for loading, user typing, and assistant thinking. Fully functional UI.

Real-time chat interface with assistant typing animation

Default dark UI with light mode toggle

Mode selector (e.g., Helper, Researcher, Therapist)

Conversation history in a left-side nav

Smart input field with command suggestions

System message configuration per session

We kept it ultra-focused, no noise, just flow.

The font gives it that “I’m talking to the future” feel, while spacing and hierarchy keep things digestible. The assistant’s mode switcher is subtle but powerful. users can shape the tone of the AI without leaving the chat. Light mode is optional, but dark mode was the default for that sleek sci-fi aesthetic.

Typing animation stuck? Double-check the message stream structure and loop triggers.

Messages not scrolling to bottom? Use scrollIntoView() after render.

API responses showing as objects? Sanitize outputs or stringify as needed.

Mode switching doesn’t affect tone? Ensure the selected system prompt is being re-injected on switch.

Slow response UX? Add “thinking…” animation + streamed token rendering to mask delays.