conda run

Run an executable in a conda environment

The conda run command allows you to run executables within a specified conda environment without manually activating it first.

Syntax

conda run [options] [--] EXECUTABLE [ARGS ...]

Arguments

EXECUTABLE [ARGS ...]

Executable name with additional arguments to be passed to the executable on invocation.

Use -- to separate conda options from the executable’s options:

conda run -n myenv -- python -v
conda run -v -n myenv -- tar -tvf file.tar

Options

-n, --name NAME

Name of the environment to run the command in.

conda run -n myenv python script.py

-p, --prefix PREFIX

Full path to the environment prefix to run the command in.

conda run -p /path/to/env python script.py

--cwd DIRECTORY

Current working directory for the command to run in. Defaults to the user’s current working directory if not specified.

conda run -n myenv --cwd /tmp python script.py

-v, --verbose

Use once for detailed output, twice for even more detailed output.

conda run -v -n myenv python script.py

-s, --no-capture-output / --live-stream

Don’t capture stdout/stderr (standard out/standard error). By default, conda run captures output and displays it after the command completes. This flag streams output in real-time.

conda run -n myenv --live-stream python script.py

--debug-wrapper-scripts

When set, where implemented, the shell wrapper scripts will use the echo command to print debugging information to stderr (standard error).

conda run --debug-wrapper-scripts -n myenv python script.py

--dev

(Deprecated in 26.9, will be removed in 27.3)

Sets CONDA_EXE to python -m conda, assuming the current working directory contains the root of conda development sources. This is mainly for use during tests where we test new conda sources against old Python versions.

This option is deprecated. Set PYTHONPATH to the conda source root instead.

Common Use Cases

Run a Python script in a specific environment

# Create an environment
conda create -y -n my-python-env python=3.11

# Run a script without activating
conda run -n my-python-env python script.py

Check Python version in an environment

conda run -n my-python-env python --version

Run command with arguments that start with dashes

Use -- to separate conda options from the executable’s options:

conda run -n myenv -- python -v script.py

Run command in specific directory

conda run -n myenv --cwd /path/to/project python test.py

Stream output in real-time

By default, conda run captures and displays output after completion. For long-running processes:

conda run -n myenv --live-stream python long_running_script.py

Run command with environment prefix

If you prefer to specify the full path:

conda run -p ~/envs/myenv python script.py

How It Works

When you run conda run:

1. Conda creates a temporary shell wrapper script
2. The script sets up the environment (similar to activation)
3. The command is executed within that environment
4. Output is captured (unless --no-capture-output is used)
5. The wrapper script is cleaned up after execution
6. The exit code of the command is returned

conda run is useful for:

• Running commands in environments without manual activation
• Scripting and automation
• CI/CD pipelines
• Running one-off commands in different environments

Exit Codes

The exit code of conda run matches the exit code of the executed command. If the command fails, conda run will also fail with the same exit code.

Example: Complete Workflow

# Create an environment with specific packages
conda create -y -n data-analysis python=3.11 pandas numpy

# Run a data processing script
conda run -n data-analysis python process_data.py

# Run with verbose output
conda run -v -n data-analysis python process_data.py

# Run and stream output in real-time
conda run -n data-analysis --live-stream python long_process.py