Vibe AI | Universal Chat Client

Vibe AI is a sleek, lightning-fast, single-file web interface for interacting with any standard OpenAI-compatible API endpoint. Whether you are running local models via Ollama/LM Studio or connecting to premium cloud providers, Vibe AI offers a perfectly polished, zero-setup frontend.

No installation, no build steps, no node_modules—just open the file in your browser.

✨ Features

💬 Universal API Support

• Bring Your Own API: Connect seamlessly to any backend that supports standard /models and /chat/completions endpoints.
• Dynamic Model Fetching: Click to fetch and cache all available models directly from your chosen provider.
• System Prompts: Full support for custom system instructions to shape the AI's behavior and personality.
• Multimodal Vision: Attach images (via URL or local File Upload) to chat with vision-capable models.

🛠️ Advanced Message Controls

• Message Actions Toolbar: Hover over (or tap on mobile) any message to reveal instant quick-actions.
• Inline Editing: Edit both user prompts and AI responses directly inside the chat bubble. Supports full Markdown re-rendering.
• Regenerate & Retry: Instantly regenerate an AI's response or resubmit a past prompt with a single click.
• Copy & Delete: Copy raw text/markdown instantly, or delete specific messages from the history context.

⚡ Peak Performance & Reliability

• Single Source of Truth Architecture: Built on a flawlessly synchronized state engine. History strictly tracks the visible UI, eliminating race conditions or duplicate contexts, even when modifying history during an active AI generation.
• Cancellable Requests: Instantly abort in-flight generations. Zero race conditions, zero UI locks, and no memory leaks.
• Bulletproof State: Your settings, selected models, and chat history are saved securely to your browser's LocalStorage with complete corruption-resistance.
• Markdown Support: Full support for GitHub Flavored Markdown (GFM), intelligent line breaks, code syntax highlighting, tables, and formatted text.

🎨 Beautiful UI / UX

• Standalone Architecture: The entire application runs perfectly from one isolated index.html file.
• Responsive & Mobile Optimized: A beautifully optimized, flexbox-driven layout. Mobile features a custom tap-to-activate logic to prevent accidental button clicks on touch screens.
• Dynamic Dark/Light Mode: Automatic theme detection based on your OS preference, complete with a manual override toggle and high-contrast element mapping.
• Auto-Resizing Input: The chat box scales comfortably as you type long, multi-line prompts.

🚀 Getting Started

Prerequisites

• A modern web browser (Chrome, Firefox, Safari, Edge).
• An active API endpoint (e.g., OpenAI, Ollama, LM Studio, Groq, Anthropic via proxy, etc.).

Installation & Usage

1. Download the index.html file from this repository.
2. Open the file in your browser (or host it locally via a simple HTTP server).
3. Open the Sidebar (☰) and configure your connection:
   • API Base URL: Enter your endpoint (e.g., http://localhost:11434/v1 or https://api.openai.com/v1).
   • API Key: Enter your key (if your provider requires authentication).
4. Click Models to populate the dropdown.
5. Select your model and start chatting!

🧩 Technologies Used

• Core: HTML5, CSS3, Vanilla JavaScript.
• Network: Axios for rock-solid HTTP requests and request cancellation.
• Formatting: Marked.js for beautiful Markdown rendering.
• Icons: FontAwesome.

⚠️ Disclaimer

• All data, chat history, and API keys are stored exclusively and securely in your browser's Local Storage.
• Clearing your browser cache or site data will wipe your chat history and settings. No data is ever sent to third parties except the API provider you explicitly configure.

📄 License

This project is open source and available under the MIT License.