nanochat

Repository: https://github.com/dustinwloring1988/nanochat

A modern, feature-rich web interface built with SvelteKit. It supports both llama.cpp (OpenAI-compatible endpoints) and Ollama backends while keeping the same clean interface, streaming UX, file handling, and conversation management.

Features

• Modern Chat Interface - Clean, responsive design with dark/light mode
• File Attachments - Support for images, text files, PDFs, and audio with rich previews and drag-and-drop support
• Conversation Management - Create, edit, branch, and search conversations
• Advanced Markdown - Code highlighting, math formulas (KaTeX), and content blocks
• Reasoning Content - Support for models with thinking blocks
• Keyboard Shortcuts - Keyboard navigation (Shift+Ctrl/Cmd+O for new chat, Shift+Ctrl/Cmdt+E for edit conversation, Shift+Ctrl/Cmdt+D for delete conversation, Ctrl/Cmd+K for search, Ctrl/Cmd+V for paste, Ctrl/Cmd+B for opening/collapsing sidebar)
• Request Tracking - Monitor processing with slots endpoint integration
• UI Testing - Storybook component library with automated tests

Development

Install dependencies:

npm install

Start the development server + Storybook:

npm run dev

This will start both the SvelteKit dev server and Storybook on port 6006.

Building

Create a production build:

npm run build

The build outputs static files to ../public directory. You can serve these with a reverse proxy in front of llama.cpp or Ollama.

Using with Ollama (local default):

• Set Provider to Ollama in Settings
• Server URL: http://localhost:11434 (default)
• Ensure CORS allows your UI origin if serving remotely

Using with llama.cpp or OpenAI-compatible servers:

• Set Provider to OpenAI-compatible
• Keep relative endpoints (default) or proxy to your server
• If your server requires a key, set API Key in Settings

Testing

Run the test suite:

# E2E tests
npm run test:e2e

# Unit tests
npm run test:unit

# UI tests
npm run test:ui

# All tests
npm run test

Architecture

• Framework: SvelteKit with Svelte 5 runes
• Components: ShadCN UI + bits-ui design system
• Database: IndexedDB with Dexie for local storage
• Build: Static adapter for deployment with llama.cpp server
• Testing: Playwright (E2E) + Vitest (unit) + Storybook (components)