Getting Started with Pixeltable

This is a step-by-step guide to setting up a local installation of Pixeltable. If you'd prefer to use a cloud notebook platform such as Kaggle or Colab, see the "Quick Start" section of the Overview instead.

You'll want to install Pixeltable in a Python virtual environment; we'll use Apache Miniconda in this guide, but any environment manager should work.

Pixeltable works with Python 3.9, 3.10, 3.11, or 3.12 running on Linux, MacOS, or Windows.

Install Pixeltable

1. Install Miniconda here: Installing Miniconda

2. Create your environment:

   • conda create --name pxt python=3.9 Any python version between 3.9 and 3.12 should work, but 3.9 is preferred for normal development to ensure compatibility.
   • conda activate pxt

3. Install Pixeltable and Jupyter inside the new environment:

   • pip install pixeltable jupyter

Create a Notebook

4. Start your Jupyter notebook server:

   • jupyter notebook

5. Create a new notebook:

   • Select "Python 3 (ipykernel)" as the kernel
   • File / New / Notebook

6. Test that everything is working by entering these commands into the notebook:

import pixeltable as pxt
pxt.init()

7. Wait a minute for Pixeltable to load; then you should see a message indicating that
   Pixeltable has successfully connected to the database.

👍

At this point, you're set up to start using pixeltable! For a tour of what it can do, a good place to start is the Pixeltable Basics tutorial.