Ullama Chat App

A modern, multi-model AI chat application built with Next.js, React, and Tailwind CSS. This project provides a beautiful chat interface for interacting with local Ollama models, supporting multiple conversations, markdown rendering, and model selection.

Features

• Chat with local Ollama models (Llama 3, Gemma, Mistral, Code Llama, Phi 3, etc.)
• Multiple conversations with dynamic topic naming
• Markdown and code rendering in responses
• Responsive, modern UI with dark mode support
• Model selection per conversation

Getting Started

1. Clone the Repository

git clone [email protected]:utshabeb/ullama.git
cd ullama

2. Install Dependencies

yarn install
# or
npm install

3. Start Ollama (if not already running)

Make sure you have Ollama installed and running locally. You can start a model with:

ollama run llama3

4. Run the Development Server

yarn dev
# or
npm run dev

Open http://localhost:3000 in your browser to use the app.

Changing the Default Model Selection

The default model for new conversations is set in src/app/page.tsx:

const [lastChosenModel, setLastChosenModel] = useState('llama3.2:latest');

To change the default, update the string (e.g., to 'mistral', 'gemma', etc.).

You can also edit the <select> options in the sidebar to add or remove available models.

---

Project Structure

• src/app/page.tsx — Main chat UI and logic
• src/app/layout.tsx — App layout and metadata
• src/app/api/ollama/route.ts — API route for communicating with Ollama
• public/ — Static assets
• tailwind.config.ts — Tailwind CSS configuration

License

MIT