How to Use Shell Commands in Google Colab 2026

Sharing is caring!

Introduction: How to Use Shell Commands in Google Colab 2026

Knowing how to use shell commands in Google Colab unlocks the full power of Linux inside your notebook. From installing packages and managing files to checking GPUs, cloning Git repos, or running scripts, Colab supports nearly everything a real terminal can do.
This guide walks you through shell basics, automation tricks, real examples, best practices, and common errors.

Run Shell Commands in Google Colab Using the ! Operator

Google Colab executes Linux commands using the ! prefix.

Examples:

!ls
!pwd
!mkdir test_folder
!rm -r test_folder

These commands run directly in Colab’s temporary Linux environment and behave like standard terminal operations.

Run Multiple Commands Using %%bash

For scripts or sequences of commands, use the %%bash magic:

%%bash
echo "Compiling program"
mkdir build
cd build
echo "Done!"

This is ideal for long workflows, loops, functions, and shell automation.

Install System Packages With apt-get

You can install Linux packages exactly as you would on Ubuntu:

!sudo apt-get update
!sudo apt-get install ffmpeg -y

Use this to install development tools, compilers, libraries, and system utilities.

Manage Files and Folders Using Shell Commands

Here are the most useful day-to-day commands:

Example:

!ls /content

How to Use Shell Commands in Google Colab 2026

Install Python Packages Using Shell Commands

Colab supports both !pip and %pip:

!pip install numpy
%pip install pandas

Important: %pip installs packages into the active runtime.
!pip can install into the wrong environment, so %pip is recommended.

Internal link: You may also like How to Debug Python Code in Google Colab on darekdari.com.

Clone GitHub Repositories in Google Colab

Clone any repo using:

!git clone https://github.com/username/repo.git

Update the repo later with:

!git pull

This is widely used in machine learning and dataset preparation.

Check System Information With Shell Commands

Colab allows you to check hardware specs easily:

!nvidia-smi
!lscpu
!free -h
!df -h

Example to check GPU model and VRAM:

!nvidia-smi --query-gpu=name,memory.total --format=csv

Create and Run Bash Scripts Inside Colab

To build and run .sh scripts:

%%bash
echo 'echo "Hello from script"' > script.sh
chmod +x script.sh

Execute it:

!./script.sh

Great for automating training loops or preprocessing pipelines.

Mix Python and Shell Commands in One Notebook

Send Python variables to shell:

filename = "data.txt"

!echo "Processing $filename"

Send shell output back to Python:

!echo 42 > number.txt

with open("number.txt") as f:
    print(int(f.read()))

This makes hybrid automation extremely powerful.

Troubleshooting and Common Mistakes

“bash: command not found”
The package isn’t installed. Install with:

!sudo apt-get install packagename

“Permission denied”
Add permissions:

!chmod +x script.sh

“No such file or directory.”
Paths changed. Re-check with:

!pwd
!ls

Package installed but Python cannot import it
Use %pip install, not !pip install.

Files disappearing after refresh
Save everything to Google Drive instead of /content.

Internal link: Learn How to Mount Google Drive in Colab on darekdari.com.

Best Practices for Using Shell Commands in Colab

• Prefer %pip and %apt for reliable installations
• Save important files to Drive
• Use %%bash for long shell scripts
• Check GPU availability before training
• Keep directories simple
• Comment your commands for clarity
• Avoid commands that alter core system files

Alternatives to Shell Commands

• Python’s subprocess module
• Jupyter magic commands
• Google Drive sidebar file manager
• GitHub Codespaces
• Google Cloud Shell for advanced workflows

Frequently Asked Questions

1. How do I run shell commands in Google Colab?
Add ! before any command, or use %%bash for multiple commands.

2. Can I run bash scripts in Colab?
Yes — create a .sh file and run it normally.

3. How do I install packages like ffmpeg in Colab?
Use:

!sudo apt-get install ffmpeg

4. Why do files disappear after I restart Colab?
The environment resets. Save to Drive to avoid losing work.

5. Can I mix Python code and shell commands together?
Yes — Colab supports variable injection both ways.

6. How do I check if a GPU is available?
Run:

!nvidia-smi

7. What’s the difference between !pip and %pip?
%pip installs correctly into the active runtime.

8. Can I use Git inside Google Colab?
Absolutely — clone repos with !git clone.

9. How do I navigate directories with shell commands?
Use !ls to inspect and avoid path mistakes.

10. Can I run advanced Linux commands in Colab?
Yes, but root-level commands may be restricted.

11. How do I delete files in Colab?
Use !rm for files and !rm -r for folders.

12. How do I fix pip packages not being detected?
Restart the runtime after installing with %pip install.

Conclusion

Understanding how to use shell commands in Google Colab gives you full control over file management, system tools, installations, and automation. Whether you’re training models, processing data, or building scripts, shell commands make your workflow faster and more professional.

For more practical Colab tutorials, visit darekdari.com.