Local AI Assistant

Local AI Assistant is a lightweight web front-end that runs entirely on your machine and chats with a local LLM through Ollama. It’s designed to answer JavaScript, SQL, and HTML questions precisely — short steps, correct code blocks, and engine/database notes when relevant.

Works on macOS and Windows with simple helper scripts; no external API keys required. Syntax highlighting and Markdown rendering are bundled for fully offline use.

What’s inside

  • Clean UI with intro box, temperature control, and “copy code” buttons.
  • Strict domain: JavaScript, SQL, and HTML answers only.
  • Local vendor assets (marked + highlight.js) for offline usage.
  • Health check badge for Ollama connectivity.

Quick Start

1) Clone the repo

git clone https://github.com/omidteimoori/local-llm-assistant.git
cd local-llm-assistant

2) Run on macOS

chmod +x scripts/run-mac.sh
./scripts/run-mac.sh

Installs Ollama via Homebrew if needed, starts ollama serve, pulls qwen2.5-coder:7b-instruct, serves the UI at http://localhost:5173, and opens your browser.

3) Run on Windows (PowerShell)

.\scripts\run-windows.ps1

Installs Ollama via winget if needed, starts ollama serve, pulls the model, serves the UI at http://localhost:5173, and opens your browser.

Prefer manual setup?
  1. Install Ollama: ollama.ai/download
  2. Run: 
    ollama serve
ollama pull qwen2.5-coder:7b-instruct
  3. Serve the frontend: 
    python3 -m http.server 5173
  4. Open http://localhost:5173

Developed by Omid Teimoori.

→ Visit GitHub Repository