How to Install Miniconda: A Step-by-Step Guide for All Platforms

Miniconda is a lightweight installer for Conda — the open-source package and environment manager used heavily in Python-based development, data science, and machine learning workflows. Unlike Anaconda, which ships with hundreds of pre-installed packages, Miniconda gives you a minimal base: just Conda itself, Python, and a handful of dependencies. From there, you build only what you need.

Understanding how the installation works — and what shapes the process — helps you avoid common pitfalls before you run a single command.

What Miniconda Actually Installs

When you install Miniconda, you're getting three core things:

• Conda — the package and environment manager
• Python — either the version bundled with the installer or one you specify later
• A base environment — an isolated space from which you create additional environments

Miniconda does not install Jupyter, NumPy, pandas, or any scientific libraries by default. Those come later, installed into specific environments you create. This is the whole point: a clean slate.

Choosing the Right Installer

Before downloading anything, two decisions matter:

1. Operating System Miniconda has distinct installers for Windows, macOS, and Linux. Each behaves differently during setup, particularly around shell integration and file paths.

2. Architecture Modern systems generally use x86-64 (64-bit) processors, but Apple Silicon Macs (M1, M2, M3 chips) require the arm64 installer. Installing the wrong architecture can cause packages to run slowly through emulation — or not at all. Check your system specs before downloading.

You can always verify the correct installer on the official Miniconda documentation page.

Installing on Windows 🖥️

1. Download the .exe installer from the official Conda site
2. Run the installer as a standard user (administrator rights are only needed if installing system-wide)
3. When prompted, choose "Just Me" unless you specifically need all users on the machine to have access
4. Leave the default installation path unless you have a reason to change it — spaces in directory names can cause issues with some tools
5. On the Advanced Options screen, you'll see a checkbox to "Add Miniconda3 to my PATH" — this is unchecked by default, which is intentional; Conda recommends using the Anaconda Prompt instead of modifying the system PATH
6. After installation, open Anaconda Prompt from the Start menu and run conda --version to confirm it's working

Installing on macOS

macOS offers two installer formats: a graphical .pkg file and a command-line .sh script. The .pkg is simpler for most users; the .sh gives more control.

Using the .sh installer:

bash Miniconda3-latest-MacOSX-arm64.sh 

Replace arm64 with x86_64 for Intel Macs. The installer will walk through a license agreement, ask where to install (default: ~/miniconda3), and then ask whether to initialize Miniconda — this modifies your shell's config file (.zshrc or .bash_profile) to make the conda command available automatically.

After installation, close and reopen your terminal, then run:

conda --version 

Installing on Linux 🐧

The process mirrors macOS with the .sh installer:

bash Miniconda3-latest-Linux-x86_64.sh 

Linux users running on ARM hardware (like Raspberry Pi or cloud instances) should use the aarch64 variant. The same shell initialization prompt applies — answering yes adds Conda to your shell startup file automatically.

Shell Initialization: The Step People Skip

One of the most common post-install issues is the conda command not being found. This almost always traces back to shell initialization not being completed.

Running conda init after installation writes activation code to your shell config. On macOS and Linux, that's typically .zshrc or .bashrc. On Windows, this is handled automatically through the Anaconda Prompt.

If you're using a non-default shell — fish, tcsh, PowerShell — Conda supports those too, but requires explicitly running conda init <shell-name>.

Verifying the Installation

Regardless of platform, the same verification steps apply:

conda --version conda info conda list 

conda info shows your active environment, platform, and Python version. conda list shows installed packages in the base environment — which should be minimal if Miniconda installed correctly.

What Varies by Setup

The installation itself is standardized, but several factors influence what comes next:

• Technical comfort level shapes whether the .sh script or graphical installer makes more sense
• Existing Python installations can create PATH conflicts, especially on macOS where a system Python already lives in /usr/bin
• Organizational or server environments may restrict where you can write files, affecting the installation directory
• Shell choice determines which config file gets modified and whether extra init steps are needed
• Whether you're using virtual machines or containers changes whether Miniconda is the right tool at all versus something like a Docker image with Conda pre-configured

Getting the base install right matters because every environment you create afterward inherits from it. The combination of your platform, shell, existing tooling, and intended use all determine which choices during installation will cause friction later — and which won't.