Desktop Apps
AI Lab Desktop

AI Lab Desktop

A standalone version of AI Lab for local data science work.

Overview

AI Lab Desktop brings the full power of Calliope’s AI-enhanced JupyterLab to your local machine. Work offline, leverage your GPU, and maintain full control of your data.

Features

JupyterLab Environment

  • Full JupyterLab interface
  • Multiple kernels (Python, R, Julia)
  • Extensions support
  • Terminal access
  • File browser

AI Integration

  • Jupyter AI chat: Converse with AI in a sidebar panel
  • Magic commands: %%ai and %ai for inline AI
  • Calliope magic: %calliope for SQL and data operations
  • Code assistance: AI-powered completion and suggestions

Data Agent

Built-in data agent capabilities:

  • Natural language SQL queries
  • RAG over your documents
  • Code execution assistance
  • Visualization generation

Installation

Download

  1. Go to calliope.ai/download
  2. Select AI Lab Desktop
  3. Choose your platform:
    • macOS (Apple Silicon)
    • macOS (Intel)
    • Windows
    • Linux

macOS Installation

  1. Download the .dmg file
  2. Open the downloaded file
  3. Drag AI Lab to Applications
  4. First launch: Right-click → Open (to bypass Gatekeeper)

Windows Installation

  1. Download the .exe installer
  2. Run the installer
  3. Follow the setup wizard
  4. Launch from Start menu

Linux Installation

AppImage:

chmod +x AILab-*.AppImage
./AILab-*.AppImage

Debian/Ubuntu:

sudo dpkg -i ailab-*.deb

First Launch

Initial Setup

  1. Launch AI Lab Desktop
  2. The setup wizard opens
  3. Configure AI provider:
    • Select provider (OpenAI, Anthropic, Ollama)
    • Enter API key (for cloud providers)
  4. Click Complete Setup

Creating Your First Notebook

  1. Click + to create new notebook
  2. Select Python kernel
  3. Start coding!

Testing AI Features

Try the AI chat:

  1. Open the Jupyter AI panel (right sidebar)
  2. Type: “Hello, can you help with data analysis?”
  3. See the AI respond

Try magic commands:

%ai explain this code
def fibonacci(n):
    return n if n < 2 else fibonacci(n-1) + fibonacci(n-2)

Configuring AI Providers

OpenAI

  1. Go to Settings → AI Provider
  2. Select OpenAI
  3. Enter your API key
  4. Choose model (gpt-4o recommended)

Anthropic

  1. Go to Settings → AI Provider
  2. Select Anthropic
  3. Enter your API key
  4. Choose model (claude-3-5-sonnet recommended)

Ollama (Local Models)

For fully offline AI:

  1. Install Ollama on your machine
  2. Pull a model: ollama pull llama3
  3. In AI Lab: Settings → AI Provider → Ollama
  4. Select from available local models

Using AI Features

Jupyter AI Chat

The sidebar chat for conversations:

You: How do I read a CSV file with pandas?

AI: Here's how to read a CSV file:
```python
import pandas as pd
df = pd.read_csv('your_file.csv')

AI Magic Commands

Cell magic (full cell):

%%ai gpt4o
Explain the pandas groupby operation with an example

Line magic (single line):

%ai summarize: df.describe()

Calliope Magic

For data operations:

%calliope list-datasources
%%calliope ask-sql mydb
What are the top 10 customers by revenue?

Working with Data

Local Files

Access files from your machine:

  • Use the file browser
  • Drag and drop into notebooks
  • Access via standard paths

Database Connections

Connect to local or remote databases:

%%calliope add-database
{
    "datasource_id": "local_db",
    "name": "Local PostgreSQL",
    "dialect": "postgresql",
    "connection_details": {
        "host": "localhost",
        "port": 5432,
        "database": "mydata",
        "user": "myuser"
    },
    "password_field": "local_db_pass",
    "password_value": "mypassword"
}

Working with Large Files

AI Lab Desktop can leverage your local resources:

  • Large datasets that exceed cloud limits
  • GPU acceleration for ML workloads
  • Local file system for big data

Extensions

Pre-installed Extensions

  • jupyterlab-git
  • jupyterlab-lsp
  • jupyter-ai
  • calliope extensions

Installing Extensions 

pip install jupyterlab-extension-name

Or via JupyterLab UI:

  1. Go to Extension Manager
  2. Search for extension
  3. Click Install

Kernels

Default Kernels

  • Python 3.11+
  • Additional kernels available

Adding Kernels

R kernel:

R -e "install.packages('IRkernel'); IRkernel::installspec()"

Julia kernel:

julia -e 'using Pkg; Pkg.add("IJulia")'

Performance Tips

For Large Datasets

  • Use chunked reading for big files
  • Leverage local SSD for caching
  • Use appropriate data types

GPU Acceleration

If you have a compatible GPU:

  1. Install CUDA toolkit (NVIDIA)
  2. Install GPU-enabled libraries:
pip install torch --index-url https://download.pytorch.org/whl/cu118

Memory Management

  • Monitor memory usage in status bar
  • Restart kernels to free memory
  • Use generators for large iterations

Offline Usage

Fully Offline Setup

  1. Configure Ollama with local models
  2. Pre-download any needed packages
  3. Prepare local data files

Required for Offline

  • AI provider: Ollama with downloaded models
  • Python packages: Pre-installed
  • Data: Local files or databases

Syncing with Cloud

Connecting to Cloud Workspace

  1. Go to Settings → Cloud Sync
  2. Sign in with Calliope account
  3. Choose sync options

What Can Sync

  • Notebooks
  • Settings
  • Extensions
  • Credentials (encrypted)

Sync Behavior

  • Two-way sync available
  • Conflict resolution options
  • Selective sync by folder

Troubleshooting

App Won’t Start

macOS: Right-click → Open for first launch Windows: Run as administrator Linux: Check AppImage permissions

AI Not Responding

  • Verify API key is correct
  • Check internet connection (cloud providers)
  • Ensure Ollama is running (local models)

Slow Performance

  • Check available RAM
  • Close unused kernels
  • Reduce model size for local AI

Extension Issues 

# Reset extension state
jupyter lab clean
jupyter lab build
IDE Desktop