Getting Started

Alpaka Desktop is a native Linux desktop client for Ollama. It connects directly to your local (or remote) Ollama instance and provides a full chat UI with streaming, model management, and advanced generation controls.

Prerequisites

Before launching Alpaka Desktop you need Ollama running:

# Install Ollama
curl -fsSL https://ollama.com/install.sh | sh

# Start the service
ollama serve

# Pull your first model
ollama pull llama3.2

Ollama listens on http://localhost:11434 by default. Alpaka Desktop connects there automatically on first launch.

Installation

See the landing page for all installation options (AppImage, APT, AUR, source build).

First Launch

1. Open Alpaka Desktop. The app icon appears in your system tray.
2. Check the host connection. A green dot next to the host name in the top bar means Ollama is reachable. If you see a red dot, go to Settings → Hosts and verify the URL (http://localhost:11434).
3. Select a model. Click the model selector in the top bar and choose a model you have pulled. If the list is empty, go to Models → Library to pull one.
4. Start chatting. Type a message and press Enter to send. Tokens stream in real time.

Interface Overview

New Chat

Launch

Models

Settings

Conversations

Today

test

hello

Explain quantum comput...

Sidebar

🦙

Type a message...

qwen3-vl:4b ▼

Message Input

• Sidebar — conversation history, organized by date. Click a conversation to open it. Use Ctrl+/ to toggle.
• Top bar — model selector, host indicator, and settings button.
• Chat area — messages render Markdown, code blocks, math (LaTeX), and thinking blocks.
• Input — Enter sends, Shift+Enter inserts a newline.

What's Next

• Chat features — streaming, thinking blocks, attachments, export
• Model management — pull, tag, create custom models
• Settings — understand the three-layer settings system