> ## Documentation Index
> Fetch the complete documentation index at: https://docs.gpuhub.com/llms.txt
> Use this file to discover all available pages before exploring further.

# JupyterLab Fails to Open

> Typically caused by dependency installation or network connection issues.

## Issue Type 1: Dependency installation

If JupyterLab fails to run due to dependency changes during package installation, log in via [SSH](/container-instance/ssh) and run `jupyter-lab` to check. If it errors, reinstall JupyterLab using `pip install jupyterlab` and restart the instance.

(Windows users can SSH via CMD or PowerShell, but may need to **manually enter the password** instead of pasting it.)

* If you see normal output after entering `jupyter-lab`, JupyterLab is working correctly.
* If you encounter errors like the ones below, JupyterLab cannot start.

<img src="https://mintcdn.com/gpuhub/yLmjMPcqL-OhC9CD/images/faq-jupyterlab-fails-1.png?fit=max&auto=format&n=yLmjMPcqL-OhC9CD&q=85&s=110cce8812b0067443191141b33dc57f" alt="" width="1080" height="642" data-path="images/faq-jupyterlab-fails-1.png" />

Solution for JupyterLab startup errors:

Reinstall JupyterLab.

* If the `pip` command is missing, run `conda install pip` first.
* If `conda` is missing, you may have deleted Miniconda, and the simplest solution is to reset the system.
  1. Uninstall the current JupyterLab: `pip uninstall jupyterlab`
  2. Reinstall JupyterLab: `pip install jupyterlab`

<img src="https://mintcdn.com/gpuhub/yLmjMPcqL-OhC9CD/images/faq-jupyterlab-fails-2.png?fit=max&auto=format&n=yLmjMPcqL-OhC9CD&q=85&s=354e27fd7cde0caebc6528f37a8da1e5" alt="" width="1080" height="162" data-path="images/faq-jupyterlab-fails-2.png" />

## Issue Type 2: Network Connection

If the first issue is ruled out, the problem is likely network-related:

1. Close all running VPN connections.
2. If access is still not possible after closing the VPN, you may need to turn the VPN back on and switch it to Direct mode.
3. If your VPN does not have a Direct mode, choose PAC mode (avoid using Global mode).
4. Try accessing with a different browser.
5. Contact customer support if the issue persists.
