Novexus

A premium AI chat interface for your local Ollama models. Built with Next.js, TypeScript, and shadcn/ui, Novexus provides a polished, feature-rich experience for running AI models entirely on your own device.

Features

• Local AI Integration — Connects to Ollama running locally on your device for private, offline AI conversations
• Tool Use — AI assistant can use built-in tools to perform real-world actions:
  • calculate — Evaluate mathematical expressions
  • clock — Get the current date and time
  • memories — Read/write persistent memories across sessions
  • files — Read, write, append, and list files in your workspace
  • run_command — Execute shell commands
  • search — Search the web via DuckDuckGo (powered by Puppeteer)
  • open_page — Open and extract content from any website URL
• Conversation Management — Create, rename, delete, and switch between multiple conversations
• Streaming Responses — Real-time streaming output with optional thinking/reasoning visibility
• Model Selection — Choose from available Ollama models on the fly
• Markdown Rendering — Full Markdown support with syntax highlighting (highlight.js), KaTeX math rendering, and GFM tables
• Customizable Settings — Configure temperature, top-p, top-k, max tokens, context length, repeat penalty, and system prompts
• Dark Theme — Beautiful dark UI with Inter & JetBrains Mono fonts
• Persistent Storage — Conversations and settings saved to localStorage
• Responsive Design — Built with Tailwind CSS v4 and shadcn/ui components

Tech Stack

• Framework: Next.js 16 (App Router, RSC)
• Language: TypeScript
• UI: React 19, shadcn/ui (New York style), Radix UI primitives
• Styling: Tailwind CSS v4, custom dark theme with oklch colors
• State: SWR for data fetching, localStorage for persistence
• Tools Runtime: Puppeteer (headless browser), JSDOM, Turndown (HTML→Markdown)
• Markdown: react-markdown with rehype/remark plugins
• Forms: react-hook-form with Zod validation
• Fonts: Inter (sans-serif), JetBrains Mono (monospace)

Prerequisites

• Node.js 18+ and pnpm/npm
• Ollama installed and running locally on http://localhost:11434
• At least one Ollama model pulled (e.g., ollama pull qwen3.5)

Getting Started

1. Clone the repository

   git clone <your-repo-url>
   cd Novexus

2. Install dependencies

   npm install
   # or
   pnpm install

3. Start Ollama (if not already running)

   ollama serve

4. Run the development server

   npm run dev
   # or
   pnpm dev

5. Open http://localhost:3000 in your browser.

Available Scripts

Command	Description
npm run dev	Start the Next.js development server
npm run build	Build the production bundle
npm run start	Start the production server
npm run lint	Run ESLint

Project Structure

Novexus/
├── app/                      # Next.js App Router
│   ├── api/                  # API routes
│   │   ├── chat/             # Chat streaming endpoint
│   │   ├── execute-tool/     # Tool execution endpoint
│   │   └── models/           # Ollama models listing
│   ├── globals.css           # Global styles & theme vars
│   ├── layout.tsx            # Root layout with fonts
│   └── page.tsx              # Main chat page
├── components/
│   ├── chat/                 # Chat UI components
│   │   ├── chat-sidebar.tsx
│   │   ├── chat-header.tsx
│   │   ├── chat-area.tsx
│   │   ├── chat-input.tsx
│   │   ├── settings-dialog.tsx
│   │   └── ...
│   ├── ui/                   # shadcn/ui components
│   └── theme-provider.tsx
├── hooks/
│   ├── use-chat-engine.ts    # Core chat engine hook
│   ├── use-mobile.ts
│   └── use-toast.ts
├── lib/
│   ├── chat-store.ts         # localStorage persistence
│   ├── tool-definitions.ts   # Tool schemas for Ollama
│   ├── tools.ts              # Server-side tool implementations
│   ├── types.ts              # TypeScript type definitions
│   └── utils.ts              # Utility functions
└── memories.json             # Persistent memory store (auto-created)

Configuration

Settings (via Settings Dialog)

Setting	Default	Description
Ollama URL	http://localhost:11434	URL of your Ollama instance
Temperature	0.7	Sampling temperature
Top P	0.9	Nucleus sampling threshold
Top K	40	Top-k sampling value
Max Tokens	4096	Maximum output tokens
Max Context	0 (unlimited)	Maximum context window size
Repeat Penalty	1.1	Repetition penalty factor
Seed	null (random)	Fixed random seed for reproducibility

System Prompt

The default system prompt can be customized via the Settings Dialog. The AI is configured as "Novexus" — a Gen Z-style assistant that runs locally on your device.