Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/conda/conda/llms.txt

Use this file to discover all available pages before exploring further.

​
What is Conda?

Conda is a cross-platform, language-agnostic package and environment manager. It helps you install, update, and manage packages and their dependencies, while keeping different projects isolated in separate environments.
Conda is developed by Anaconda, Inc. and is open source. Visit the official conda website for more information.

​
Packages vs Environments

​
Packages

A package is a bundled archive file containing:
  • Compiled software libraries and executables
  • Configuration files
  • Metadata about dependencies and versions
  • Package information in JSON format
Conda supports two package formats:
  • .conda - The modern, faster package format (v2)
  • .tar.bz2 - The legacy package format (v1)
# Install a single package
conda install numpy

# Install specific version
conda install python=3.11

# Install multiple packages
conda install scipy curl wheel

​
Environments

An environment is an isolated directory containing:
  • A specific collection of packages
  • Its own Python/R interpreter (if applicable)
  • All dependencies for those packages
  • Configuration and metadata in the conda-meta/ directory
Environments allow you to:
  • Keep dependencies for different projects separate
  • Use different versions of packages for different projects
  • Avoid conflicts between package requirements
The default environment is called base (defined as ROOT_ENV_NAME in the source). This is a reserved environment name along with root.

​
Channels and Repositories

​
What are Channels?

A channel is a repository location where conda looks for packages. Channels are URLs that point to directories containing package files and metadata.

​
Default Channels

Conda uses different default channels depending on your platform: 
DEFAULT_CHANNELS_UNIX = (
    "https://repo.anaconda.com/pkgs/main",
    "https://repo.anaconda.com/pkgs/r",
)
The default channel alias is https://conda.anaconda.org.

​
Channel Priority

Conda uses a channel priority system to determine which package to install when multiple channels have the same package:
  • Strict: Install packages only from the highest priority channel that has them
  • Flexible: Consider version and build information across all channels
  • Disabled: No channel priority enforcement
class ChannelPriority(ValueEnum):
    STRICT = "strict"
    FLEXIBLE = "flexible"
    DISABLED = "disabled"
The maximum channel priority value is 10000, reserved for packages that need to be removed or have highest priority.

​
Prefix and Environment Paths

​
What is a Prefix?

A prefix is the filesystem path where an environment is installed. Every conda environment has a unique prefix directory. 
# Default base environment
/opt/anaconda3

# Named environment
/opt/anaconda3/envs/myproject

# Custom location
/home/user/projects/ml_project/.conda

​
Prefix Structure

Every conda prefix contains: 
prefix/
├── bin/              # Executables and scripts
├── lib/              # Libraries
├── include/          # Header files
├── conda-meta/       # Environment metadata
│   ├── history       # Installation history (PREFIX_MAGIC_FILE)
│   ├── state         # Environment state
│   ├── pinned        # Pinned packages
│   ├── frozen        # Frozen environment marker
│   └── created_at    # Creation timestamp
├── etc/
│   └── conda/
│       └── env_vars.d/  # Environment variables
└── share/            # Shared data files

​
Environment Variables

Conda environments can define custom environment variables in the etc/conda/env_vars.d/ directory. 
# Value to unset a variable
CONDA_ENV_VARS_UNSET_VAR = "***unset***"

# Reserved variables (cannot be modified)
RESERVED_ENV_VARS = ("PATH",)

​
Prefix Placeholder

During package builds, conda uses a prefix placeholder for binary patching: 
PREFIX_PLACEHOLDER = "/opt/anaconda1anaconda2anaconda3"
This placeholder is replaced with the actual prefix path during installation.

​
Supported Platforms

Conda supports multiple platforms and architectures: 
PLATFORMS = (
    "emscripten-wasm32",
    "wasi-wasm32",
    "freebsd-64",
    "linux-32",
    "linux-64",
    "linux-aarch64",
    "linux-armv6l",
    "linux-armv7l",
    "linux-ppc64",
    "linux-ppc64le",
    "linux-riscv64",
    "linux-s390x",
    "osx-64",
    "osx-arm64",
    "win-32",
    "win-64",
    "win-arm64",
    "zos-z",
)
Each platform has a noarch subdirectory for platform-independent packages.

​
Configuration Files

Conda searches for configuration files (.condarc) in multiple locations: 
SEARCH_PATH = (
    "/etc/conda/.condarc",
    "/etc/conda/condarc",
    "/etc/conda/condarc.d/",
    "/var/lib/conda/.condarc",
    "/var/lib/conda/condarc",
    "/var/lib/conda/condarc.d/",
    "$CONDA_ROOT/.condarc",
    "$CONDA_ROOT/condarc",
    "$CONDA_ROOT/condarc.d/",
    "$XDG_CONFIG_HOME/conda/.condarc",
    "$XDG_CONFIG_HOME/conda/condarc",
    "$XDG_CONFIG_HOME/conda/condarc.d/",
    "~/.config/conda/.condarc",
    "~/.config/conda/condarc",
    "~/.config/conda/condarc.d/",
    "~/.conda/.condarc",
    "~/.conda/condarc",
    "~/.conda/condarc.d/",
    "~/.condarc",
    "$CONDA_PREFIX/.condarc",
    "$CONDA_PREFIX/condarc",
    "$CONDA_PREFIX/condarc.d/",
    "$CONDARC",
)
Configuration files are loaded in order, with later files overriding earlier ones. The $CONDARC environment variable takes highest precedence.

​
Next Steps

Environment Management

Learn how to create and manage conda environments

Package Management

Install, update, and remove packages

Channel Management

Configure and manage package channels