Flexible and secure .env loader for Python projects with support for environment switching, nested project structures, and external configuration directories.

• Overview
• Historical Context
• Features
• Security and Best Practices
• Installation
• Python Compatibility
• Usage
  • Basic Usage
  • Advanced Usage
  • Dry-run Usage
  • Environment Variables
  • Typical Directory Structure
• .env Resolution Rules and Precedence
• Use Cases
• License
• Contributing
• Acknowledgments
• Changelog

dotenv-loader (Git, PyPI) provides a flexible yet straightforward way to load environment variables from .env files, tailored specifically for Python applications, including Django, Flask, Celery, FastCGI, WSGI, CLI tools, and more. It supports hierarchical configuration, making it extremely convenient to manage multiple environments (development, testing, production) without manually switching .env files or cluttering code with environment-specific logic.

Managing environment-specific settings is crucial in modern software development. The excellent python-dotenv package provides robust and versatile .env file loading functionality, but it lacks flexible mechanisms for dynamically switching configurations across multiple deployment environments or nested project structures.

dotenv-loader is built upon python-dotenv, enhancing it specifically to address the following practical challenges:

• Dynamically switch environments without modifying code or manually managing environment variables.
• Support flexible directory structures, such as monorepos or Django sub-applications.
• Provide clear, hierarchical separation of environment-specific configurations to improve clarity and reduce human error.

While dotenv-loader specializes in dynamic and hierarchical environment selection tailored for specific deployment scenarios, python-dotenv offers broader and more general-purpose capabilities. Users whose needs aren't fully met by dotenv-loader are strongly encouraged to leverage python-dotenv's comprehensive functionality directly.

• Hierarchical and prioritized .env file search: dotenv-loader follows a clear and intuitive priority order:

  1. Explicit path provided via DOTENV environment variable.
  2. Configuration directory (DOTCONFIG_ROOT) with customizable subdirectories per project and environment stage (DOTSTAGE).
  3. Automatic fallback to .env file located directly in the project root.

• Dynamic project and stage selection: Quickly switch configurations by setting the DOTPROJECT and DOTSTAGE environment variables, allowing effortless toggling between multiple environments or projects.

• Customizable file names: Use custom .env filenames to further separate and manage your configurations.

Storing .env files separately in a dedicated configuration directory (DOTCONFIG_ROOT) outside your project source tree is a secure and recommended best practice. This approach significantly reduces the risk of accidentally leaking sensitive information (such as API keys, database credentials, etc.) during backups, version control operations, or file transfers. By keeping secrets separate from your codebase, dotenv-loader helps enforce a clear boundary between configuration and code, enhancing security and compliance.

pip install dotenv-loader

This package is tested on the following Python versions via GitHub Actions:

Automated tests are run on every push and pull request to ensure consistent support.

By default, dotenv-loader automatically identifies the .env file from the current project's root directory:

import dotenv_loader

dotenv_loader.load_env()  # Locate the resolved .env file and populate os.environ
                          # with its variables

from dotenv_loader import load_env

# Load environment with custom default settings:
load_env(
    project='mega-project',           # - explicitly set project name
    stage='production',               # - explicitly set stage name
    dotenv='./dir/.env.test'          # - explicitly set .env file 
    config_root='~/custom-configs',   # - custom config directory
    steps_to_project_root=1,          # - how many directories up to look for project root
    default_env_filename='custom.env',# - change the default '.env' name to you name
    override=True,                    # - whether to overwrite existing values in os.environ
    dry_run=False                     # - if True, only return the resolved .env path
                                      #   without loading it
)

Use dry-run mode when you want to inspect or parse the .env file manually (e.g., without modifying the environment):

from dotenv_loader import load_env
from dotenv import dotenv_values

env_file = load_env(dry_run=True)  # Return the resolved .env file path without applying
                                   # it to os.environ 
if env_file:
   config = dotenv_values(env_file)  # Load variables into a dict without affecting the
                                     # environment:
                                     # config = {"USER": "foo", "EMAIL": "foo@example.org"}
else:
  raise FileNotFoundError(".env file was not found!")

The load_env() function accepts the following optional parameters:

⚠️ Note: Each parameter is optional and first four parameters (project, stage, dotenv and config_root) can also be controlled via environment variables: DOTPROJECT, DOTSTAGE, DOTENV, and DOTCONFIG_ROOT, respectively.

You can control the behavior of dotenv-loader using the following environment variables:

• If a full file path is given, it overrides all other options. If the file is not found, a FileNotFoundError is raised.
• If a directory path is given, the loader will look for an environment file in that directory, based on default_env_filename and DOTSTAGE (or fallback stage).

Examples:

DOTENV=/home/user/.env.custom python manage.py
# Uses this exact file; raises an error if not found
 
DOTENV=~/myconfigs/myproject python manage.py
DOTSTAGE=prod
# Loads ~/myconfigs/myproject/.env.prod
 
DOTENV=~/configs/project python manage.py  # calling load_env(stage='local')
# Loads ~/configs/project/.env.local

DOTPROJECT=test python manage.py
# Loads: ~/.config/python-projects/test/.env

DOTSTAGE=staging python manage.py
# Loads: ~/.config/python-projects/myproject/.env.staging

DOTCONFIG_ROOT=~/myconfigs python manage.py
# Loads: ~/myconfigs/myproject/.env

DOTVERBOSE=1 python manage.py
# Output: Use DOTENV file from: /home/user/.config/python-projects/projectname/.env 

⚠️ Note
Unlike other dotenv-loader variables, DOTVERBOSE doesn't influence the selection of the .env file — it only controls whether its path is printed to stdout.
This makes it safe and convenient to define DOTVERBOSE inside your .env file (e.g. during development) to always see which file was loaded.
Supported truthy values (case-insensitive): '1', 'true', 'yes', 'on', 'ja'.
The DOTVERBOSE environment variable takes precedence over the value defined in the .env file.

~/.config/python-projects/
└── myproject/
    ├── .env          # Default configuration (typically a symlink 
    │                 # to .env.prod)
    │ 
    ├── .env.prod     # Production configuration. Use explicitly with:
    │                 # DOTSTAGE=prod python manage.py
    │ 
    ├── .env.staging  # Staging configuration. Use explicitly with: 
    │                 # DOTSTAGE=staging python manage.py
    │ 
    └── .env.test     # Testing configuration. Use explicitly with: 
                      # DOTSTAGE=test python manage.py

myproject/
└── manage.py  # By default, loads ~/.config/python-projects/myproject/.env
    .env       # Used only if no .env.* files are found in 
               # ~/.config/python-projects/myproject

dotenv-loader uses a deterministic and secure resolution strategy when selecting the appropriate .env file to load. The logic ensures maximum flexibility while maintaining clarity and safety.

1. Environment Variables Take Precedence

   The following environment variables override their corresponding load_env() arguments:

   • DOTENV overrides dotenv
   • DOTPROJECT overrides project
   • DOTSTAGE overrides stage
   • DOTCONFIG_ROOT overrides config_root

2. Relative Paths Are Context-Aware

   • Paths defined in environment variables (e.g., DOTENV, DOTCONFIG_ROOT) are resolved relative to the current working directory (PWD, as seen with pwd).

   • Paths passed directly to load_env() (e.g., dotenv, config_root) are resolved relative to the calling script's location, adjusted by steps_to_project_root.

   For example, if manage.py is at ~/projects/proj1/app/manage.py and steps_to_project_root=1, then project_root is considered to be ~/projects/proj1.

3. Project and Stage Names Must Be Basenames

   Both DOTPROJECT/project and DOTSTAGE/stage must not include slashes. They are treated strictly as simple names (i.e., Path(name).name).

4. Highest Priority: Explicit .env Path or Directory

   If either DOTENV or dotenv is defined, it takes priority over all other resolution logic.

   • If the value is a full path to a file, that file is loaded directly. If it doesn't exist, a FileNotFoundError is raised and no fallback is attempted.
   • If the value is a directory path, the loader will only search within that directory, constructing the target filename as [default_env_filename][.stage], where stage comes from DOTSTAGE or the stage argument.

5. Project Name Determination

   The project name is determined as follows:

   • If DOTPROJECT or project is defined, its basename is used
   • Otherwise, it defaults to the basename of the computed project_root directory

   ⚠️ Note: The project_root is computed relative to the file that directly calls load_env(), using the steps_to_project_root parameter.

   • If steps_to_project_root=0 (default), project_root is the directory containing the calling file
   • If steps_to_project_root=1, it's the parent of that directory, and so on

   For example:

   If load_env() is called from ~/projects/proj1/app/manage.py and steps_to_project_root=1, then project_root = ~/projects/proj1, and the fallback project name is proj1.

6. .env Filename Construction

   The .env filename is constructed as:

   "[default_env_filename][[.]STAGE]"

   where:

   • default_env_filename is .env by default
   • STAGE comes from DOTSTAGEor stage (if defined)

7. Primary Search Location

   If no explicit file path is provided via DOTENV/dotenv, the loader checks:

   [DOTCONFIG_ROOT | config_root] / [DOTPROJECT | project] / [default_env_filename][.stage]

8. Fallback Location

   If the file is not found in the config directory, a fallback search is performed in the computed project root: [project_root] / [default_env_filename][.stage]

9. Error Handling If no valid .env file is found after all resolution attempts, a FileNotFoundError is raised. The error message includes a list of all tried paths for easier debugging.

dotenv-loader is especially useful when:

• Deploying applications to multiple environments (development, testing, staging, production).
• Managing complex directory structures (monorepos or multi-app Django projects).
• Simplifying CI/CD workflows by dynamically selecting environment configurations.

MIT License

We welcome contributions from the community! Please see CONTRIBUTING for details on how to get started.

This project was created in collaboration with ChatGPT (OpenAI), utilizing the GPT-4o, GPT-4.5, and GPT-3 models.

For detailed release notes, see CHANGELOG.

By clearly managing your environment variables and enabling dynamic configuration switching, dotenv-loader helps you streamline your deployment and development workflows, reduce errors, and maintain cleaner, more maintainable code.