Verity — Grounded AI Platform Hero: live preview
Live preview: a real build of this exact prompt, recorded at 1280×800.
Hero section component (above the fold)React 18 + TypeScript + Vite + Tailwind CSS 3 + lucide-react~300 lines of code

Warm-black hero where a question sits above a streaming answer; citation chips pop in as each sentence finishes and the matching source cards light up in lime.

FontsGeist (400, 500, 600) + Geist Mono (400, 500)
PaletteWarm black #0E0E0C · Panel #171713 · Off-white #F2F0E9 · Muted #A7A59A · Acid lime #D4F26B
SaaSPremium

Verity — Grounded AI Platform Hero

A dark, centred AI-platform hero whose answer card streams a response sentence by sentence, attaching a numbered citation to each one and lighting its source in a side rail. Built for RAG, enterprise AI and any product where answers need to be checkable.

aisaasdarkcitationsragstreaming
Full prompt hidden. Copy it to use in your AI tool

Create a free account to copy 3 premium prompts. No card needed

How to use
  1. 01
    Copy the prompt
    Click the button above to copy the full prompt to your clipboard.
  2. 02
    Paste into your AI tool
    Works in Claude, ChatGPT, Cursor, v0, Lovable, or Bolt.
  3. 03
    Swap in your brand details
    Replace the brand name, copy, and image URLs with your own.
  4. 04
    Drop the code into your project
    The output is a self-contained React component. Paste it straight in.

What it generates

  • ✓Centred hero with an eyebrow, a two-line headline, a supporting paragraph and a lime primary CTA
  • ✓Answer card with a query bar, a token-by-token streaming answer and a blinking caret
  • ✓Numbered citation chips that pop in the moment each sentence completes
  • ✓Source rail with four retrieved documents that light up as they are cited, with one unused source left dim
  • ✓Grounded meter and a latency and token readout in the card footer
  • ✓Soft lime glow behind the hero and a three-stat band below the fold
  • ✓Clickable citation chips that jump to their source card
  • ✓Reduced-motion fallback that shows the finished answer with every source lit

Best for

  • →RAG and enterprise AI platforms
  • →Developer tools for building AI features
  • →Document Q&A and knowledge-base products
  • →Compliance-sensitive AI products that must show their sources

Not suitable for

  • ✕Full multi-page websites (hero only)
  • ✕Backend or database logic
  • ✕Non-React frameworks (Vue, Svelte, etc.)

Written for

  • ✓ClaudePreview built with Claude
  • ✓ChatGPTPaste into a chat
  • ✓CursorBuild inside your repo
  • ✓v0 by VercelComponent generator
  • ✓LovableApp builder
  • ✓BoltIn-browser builder

Licence

Commercial use included. Use in client projects, products, and paid work. No attribution required.

Common questions

Does this generate a full website?
No. This prompt generates a single hero section (the above-the-fold area). It includes the navigation, hero layout, and any supporting elements like a marquee or footer strip, but it is not a full multi-page site.
What do I need to run the output?
A React project with Tailwind CSS 3. The prompt uses React 18 + TypeScript + Vite + Tailwind CSS 3 + lucide-react. No additional packages are required unless specified in the prompt.
Can I customise the brand, colours, and copy?
Yes. The prompt uses a demo brand as a concrete example. Simply swap in your own brand name, headline, body copy, image URLs, and colour values when prompted by your AI tool.
Will the output look exactly like the preview?
The preview is a real build of this exact prompt, made with Claude. AI output always varies a little, and other tools may interpret small details such as spacing or timing differently, but the layout, copy and signature interaction are all specified in the prompt.
Can I use this for client work?
Yes. A commercial licence is included with every copy. You can use the output in client projects, paid products, and agencies without attribution.
What if the output has bugs?
Paste the error back into your AI tool and ask it to fix it. Most issues resolve in one or two follow-up messages. You can also re-paste the original prompt to start fresh.