Getting Started


Python 3.x is required
$ pip install dynaconf
Default installation supports .toml, .py and .json file formats and also environment variables (.env supported) - to support YAML add pip install dynaconf[yaml] or pip install dynaconf[all]


Accessing config variables in your Python application

In your Python program wherever you need to access a settings variable you use the canonical object from dynaconf import settings:

Example of program to connect to some database

from some.db import Client

from dynaconf import settings

conn = Client(
    username=settings.USERNAME,             # attribute style access
    password=settings.get('PASSWORD'),      # dict get style access
    port=settings['PORT'],                  # dict item style access
    timeout=settings.as_int('TIMEOUT'),     # Forcing casting if needed
    host=settings.get('HOST', 'localhost')  # Providing defaults

Understanding the settings

Dynaconf aims to have a flexible and usable configuration system. Your applications can be configured via a configuration files, through environment variables, or both. Configurations can be separated into environments: [default], [development], [staging], [testing] and [production]. The working environment is switched via an environment variable.

But this is all optional you can of course follow strictly the 12 factor app guide, have your configuration coming only from environment variables and provide files only to store [default] values. (take also a look on how to add a dynaconf validation file to your project).

Sensitive data like tokens, secret keys and password can be stored in .secrets.* files and/or external storages like Redis or vault secrets server.

Besides the built-in optional support to Redis as settings storage dynaconf allows you to create Custom Loaders and store the data wherever you want e.g: databases, memory storages, other file formats, nosql databases etc.

Working environments

At any point in time, your application is operating in a given configuration environment. By default there are four such environments:

  • [development] (selected by default)
  • [staging]
  • [testing]
  • [production]
  • [{custom}] <– You can create named environments that you need
There is also the pseudo-envs [default] to provide comprehensive default values and [global] to provide global values to override in any other environment.

Without any action, your applications by default run in the [development] environment. The environment can be changed via the ENV_FOR_DYNACONF environment variable. For example, to launch an application in the [staging] environment, we can run:

export ENV_FOR_DYNACONF=staging


ENV_FOR_DYNACONF=staging python
NOTE: When using Flask Extension the environment can be changed via FLASK_ENV variable and for Django Extension you can use DJANGO_ENV.

The settings files

NOTE: The settings files are optional. If it is not present, only the values from environment variables and enabled external loaders are used (.env file is also supported).

Dynaconf will search for the settings files defined in SETTINGS_FILE_FOR_DYNACONF which by default is a list containing combinations of settings.{py|toml|json|ini|yaml} and .secrets.{py|toml|json|ini|yaml} and dynaconf will try to find each one of those combinations, optionally it is possible to configure it to a different set of files e.g: export SETTINGS_FILE_FOR_DYNACONF='["myfilename.toml", "another.json"]', this value contains a list of relative or absolute paths, can be a toml-like list or a comma/semicolon separated string and can be exported to envvars, write to .env file or passed directly to Dynaconf instance.

IMPORTANT: Dynaconf by default reads settings files using utf-8 encoding, if you have settings files written in other encoding please set ENCODING_FOR_DYNACONF environment variable.

See more details in configuration

Settings files location

To find the files defined in SETTINGS_FILE_FOR_DYNACONF the search will start at the path defined in ROOT_PATH_FOR_DYNACONF (if defined), then will recursively walk to its root and then will try the folder where the called program is located and then it will recursively try its parent directories until the root parent is reached which can be File System / or the current working dir then finally will try the current working directory as the last option.

NOTE: If by any reason you need Dynaconf to first look at the current working dir you can customize the ROOT_PATH_FOR_DYNACONF via environment variable or by creating a custom settings object

Some people prefer to put settings in a sub-folder so for each of the paths it will also search in a relative folder called config/.

And for each file dynaconf will also try to load a .local. file, for example, if you have a settings.toml after loading it Dynaconf will also try to find a settings.local.toml if exists.

Dynaconf will stop searching on the first match for each file and if no file is found it will fail silently unless SILENT_ERRORS_FOR_DYNACONF=false is exported.

Illustrative Example

New in 2.0.0

If your program has the following structure:

|_ myprogram/
   |_ src/
         # from dynaconf import settings
         # print(settings.NAME)
         # print(settings.PASSWORD)
         # print(settings.FOO)
   |_ config
      |_ settings.toml
         # [default]
         # name = "Jon Doe"
   |_ settings.local.toml
      # [default]
      # name = "Oscar Wilde"
   |_ .env
   |_ .secrets.toml
      # [default]
      # password = "Utopi@"

And you call it from myprogram working dir.

cd myprogram
python src/

What happens is:

NOTE: The behavior explained here is valid only for the above file structure, other arrangements are possible and depending on how folders are organized dynaconf can behave differently.
  1. does from dynaconf import settings
    • Only the .env file will be searched, other settings are lazy evaluated.
    • .env will be searched starting on myprogram/src/.env
    • if not found then myprogram/src/config/.env
    • if not found then myprogram/.env actually found here so stops searching
    • if not found then myprogram/config/.env
    • It will load all values from .env to the environment variables and create the instance of settings
  2. does the first access to a settings on print(settings.NAME)
    • Dynaconf will execute the loaders defined in CORE_LOADERS and LOADERS, it will initialize the settings object and start the file search.
    • settings.{py|toml|json|ini|yaml} will be searched on myprogram/src/
    • if not found then myprogram/src/config
    • if not found then myprogram/
    • if not found then myprogram/config settings.toml actually found here so stops searching for toml
    • It will load all the values defined in the settings.toml
    • It will continue to look all the other files e.g: settings.json, settings.ini, settings.yaml etc.
    • Then
    • It will search for .secrets.{py|toml|json|ini|yaml} on myprogram/src/
    • if not found then myprogram/src/config
    • if not found then myprogram/ .secrets.toml actually found here so stops searching for toml
    • if not found then myprogram/config
    • It will load all the values defined in .secrets.toml (if filename is *.secret.* values are hidden on logs)
    • It will continue to look all the other files e.g: .secrets.json, .secrets.ini, .secrets.yaml etc.
    • Then
    • It will iterate the list of loaded files containing [settings.toml, .secrets.toml] and for each of them it will also try to find a settings.local.toml (found in myprogram/settings.local.toml) and a .secrets.local.toml using the same search tree until it is found or it will skip if not found.
    • Then
    • It will execute external loaders like Redis and Vault if enabled.
    • It will execute custom loaders if configured.
    • Then finally
    • It will read all environment variables prefixed with DYNACONF_ and load its values, in our example it loads FOO='BAR' from .env file.
  3. does second access to a settings on print(settings.PASSWORD)
    • All the loaders, loaded files, config options and vars are now cached no loading has been executed.
    • Only if settings.get_fresh('PASSWORD') is used, dynaconf will force a re-load of everything to ensure the fresh value.
    • Also if settings.using_env|from_env or ENV_FOR_DYNACONF switched, e.g: from [development] to [staging], then re-load happens.
    • It is also possible to explicitly force a load or reload.
  4. Complete program output is:
Oscar Wilde
TIP: If you add DEBUG_LEVEL_FOR_DYNACONF=DEBUG on .env or export this variable then you can follow the dynaconf loading process.

Loading order

Dynaconf loads file in a overriding cascade loading order using the predefined order:

  1. First the environment variables (and .env file) to read for configuration options
  2. Then the paths provided in PRELOAD_FOR_DYNACONF using all enabled loaders.
  3. Then the files defined in SETTINGS_FILE_FOR_DYNACONF using all enabled loaders.
    • Files containing .local. in its name will be loaded at the end. e.g: settings.local.yaml
  4. Then contents of SECRETS_FOR_DYNACONF envvar filename if defined (useful for jenkins and other CI)
  5. Then the loaders defined in LOADERS_FOR_DYNACONF
    • Redis if enabled by REDIS_FOR_DYNACONF=1
    • Vault if enabled by Vault_FOR_DYNACONF=1
    • Custom loaders if any added
    • Environment variables loader will be the last always
  6. If there is any DYNACONF_INCLUDE key found or INCLUDES_FOR_DYNACONF env vars this will be loaded.

The order can be changed by overriding the SETTINGS_FILE_FOR_DYNACONF the CORE_LOADERS_FOR_DYNACONF and LOADERS_FOR_DYNACONF variables.

NOTE: Dynaconf works in an layered override mode based on the above order, so if you have multiple file formats with conflicting keys defined, the precedence will be based on the loading order. If you dont want to have values like lists and dicts overwritten take a look on how to merge existing values

Local configuration files and merging to existing data

New in 2.2.0

This feature is useful for maintaining a shared set of config files for a team, while still allowing for local configuration.

Any file matched by the glob *.local.* will be read at the end of file loading order. So it is possible to have local settings files that are for example not committed to the version controlled repository. (e:g add **/*.local* to your .gitignore)

So if you have settings.toml Dynaconf will load it and after all will also try to load a file named settings.local.toml if it does exist. And the same applies to all the other supported extensions settings.local.{py,json,yaml,toml,ini,cfg}


# settings.toml        # <-- 1st loaded
colors = ["green", "blue"]
parameters = {enabled=true, number=42}

# .secrets.toml        # <-- 2nd loaded  (overrides previous existing vars)
password = 1234

# settings.local.toml  # <-- 3rd loaded  (overrides previous existing vars)
colors = ["pink"]
parameters = {enabled=false}
password = 9999

So with above the values will be:

settings.COLORS == ["pink"]
settings.PARAMETERS == {"enabled": False}
settings.PASSWORD == 9999

For each loaded file dynaconf will override previous existing keys so if you want to append new values to existing variables you can use 3 strategies.

Mark the local file to be entirely merged

New in 2.2.0
# settings.local.toml
dynaconf_merge = true
colors = ["pink"]
parameters = {enabled=false}

By adding dynaconf_merge to the top root of the file mark entire file to be merged.

And then the values will be updated in to existing data structures.

settings.COLORS == ["pink", "green", "blue"]
settings.PARAMETERS == {"enabled": False, "number": 42}
settings.PASSWORD == 9999

You can also mark a single env like [development] to be merged.

# settings.local.toml
dynaconf_merge = true
colors = ["pink"]
parameters = {enabled=false}

dynaconf merge token

# settings.local.toml
colors = ["pink", "dynaconf_merge"]
parameters = {enabled=false, dynaconf_merge=true}

By adding dynaconf_merge to a list or dict marks it as a merge candidate.

And then the values will be updated in to existing data structures.

settings.COLORS == ["pink", "green", "blue"]
settings.PARAMETERS == {"enabled": False, "number": 42}
settings.PASSWORD == 9999
New in 2.2.0

And it also works having dynaconf_merge as dict keys holding the value to be merged.

# settings.local.toml
dynaconf_merge = ["pink"]  # <-- value ["pink"] will be merged in to existing colors

dynaconf_merge = {enabled=false}

Dunder merging for nested structures

For nested structures the recommendation is to use dunder merging because it it easier to read and also it has no limitations in terms of nesting levels.

# settings.local.yaml
parameters__enabled = false

The use of __ to denote nested level will ensure the key is merged with existing values read more in merging existing values.

Global merge


or put it in your .env file then Dynaconf will automatically merge all existing variables.

BEWARE: Using MERGE_ENABLED_FOR_DYNACONF can lead to unexpected results because you do not have granular control of what is being merged or overwritten so the recommendation is to use other options.

Settings File Formats

The recommended file format is TOML but you can choose to use any of .{py|toml|json|ini|yaml}.

The file must be a series of sections, at least one for [default], optionally one for each [environment], and an optional [global] section. Each section contains key-value pairs corresponding to configuration parameters for that [environment]. If a configuration parameter is missing, the value from [default] is used. The following is a complete settings.toml file, where every standard configuration parameter is specified within the [default] section:

NOTE: if the file format choosen is .py as it does not support sections you can create multiple files like for [default],, and ATTENTION: using .py is not recommended for configuration - prefer to use static files like TOML!
username = "admin"
port = 5000
host = "localhost"
message = "default message"
value = "default value"

username = "devuser"

host = ""

host = ""

host = ""

value = "this value is set for custom [awesomeenv]"

message = "This value overrides message of default and other envs"

The [global] pseudo-environment can be used to set and/or override configuration parameters globally. A parameter defined in a [global] section sets, or overrides if already present, that parameter in every environment.

IMPORTANT: the environments and pseudo envs such as [global], ['default'] affects only the current file, it means that a value in [global] will override values defined only on that file or previous loaded files, if in another file the value is reloaded then the global values is overwritten. Dynaconf supports multiple file formats but the recommendation is not to mix them, choose a format and stick with it.

For example, given the following settings.toml file, the value of address will be “” in every environment:

address = ""

address = "localhost"

address = ""
NOTE: The [env] name and first level variables are case insensitive as internally dynaconf will always use upper case, that means [development] and [DEVELOPMENT] are equivalent and address and ADDRESS are also equivalent. But the recommendation is to always use lower case in files and always use upper case in env vars and .py files (This rule does not apply for inner data structures as dictionaries and arrays).

Supported file formats

By default toml is the recommended format to store your configuration, however you can switch to a different supported format.

# If you wish to include support for more sources
pip3 install dynaconf[yaml|ini|redis|vault]

# for a complete installation
pip3 install dynaconf[all]

Once the support is installed no extra configuration is needed to load data from those files.

If you need a different file format take a look on how to extend dynaconf writing a custom loader

Additional secrets file (for CI, jenkins etc.)

It is common to have an extra secrets file that is available only when running on specific CI environment like Jenkins, usually there will be an environment variable pointing to the file.

On Jenkins it is done on job settings by exporting the secrets information.

Dynaconf can handle this via SECRETS_FOR_DYNACONF environment variable.


export SECRETS_FOR_DYNACONF=/path/to/settings.toml{json|py|ini|yaml}

If that variable exists in your environment then Dynaconf will also load it.

Including files inside files

Sometimes you have multiple fragments of settings in different files, dynaconf allow easy merging of those files via dynaconf_include.



plugin_specific_variable = 'value for development'

and even mixing different formats:plugin2.yaml

  plugin_specific_variable: 'value for production'

Then it can be merged on main settings.toml file via dynaconf_include


dynaconf_include = ["plugin1.toml", "plugin2.yaml"]
DEBUG = false
PORT = 6666

A settings file can include a dynaconf_include stanza, whose exact syntax will depend on the type of settings file (json, yaml, toml, etc) being used:

dynaconf_include = ["/absolute/path/to/plugin1.toml", "relative/path/to/plugin2.toml"]
DEBUG = false

When loaded, the files located at the (relative or absolute) paths in the dynaconf_include key will be parsed, in order, and override any base settings that may exist in your current configuration.

The paths can be relative to the base settings.(toml|yaml|json|ini|py) file, or can be absolute paths.

The idea here is that plugins or extensions for whatever framework or architecture you are using can provide their own configuration values when necessary.

It is also possible to specify glob-based patterns:

dynaconf_include = ["configurations/*.toml"]
DEBUG = false

Currently, only a single level of includes is permitted to keep things simple and straightforward.

Including via environment variable

It is also possible to setup includes using environment variable.

# A glob pattern
export INCLUDES_FOR_DYNACONF='/etc/myprogram/conf.d/*.toml'
# a single path
export INCLUDES_FOR_DYNACONF='/path/to/file.yaml'
# multiple files
export INCLUDES_FOR_DYNACONF='/path/to/file.yaml;/other/path/to/file.toml'

Programmatically loading a settings file

from dynaconf import settings
settings.load_file(path="/path/to/file.toml")  # list or `;/,` separated allowed
NOTE: programmatically loaded file is not persisted, once env is changed via setenv|ugin_env, or a reload or configure is invoked it will be cleaned, to persist it needs to go to INCLUDES_FOR_DYNACONF variable or you need to load it programmatically again.

Template substitutions

Dynaconf has 2 tokens to enable string substitutions @format and @jinja.

@format token

Dynaconf allows template substitutions for strings values, by using the @format token prefix and including placeholders accepted by Python’s str.format method Dynaconf will call it lazily upon access time.

The call will be like:

"<YOURVALUE>".format(env=os.environ, this=dynaconf.settings)

So in your string you can refer to environment variables via env object, and also to variables defined int the settings object itself via this reference. It is lazily evaluated on access it will use the final value for a settings regardless the order of load.


export PROGRAM_NAME=calculator


DB_NAME = "mydb.db"

DB_PATH = "@format {env[HOME]}/{this.current_env}/{env[PROGRAM_NAME]}/{this.DB_NAME}"
  • {env[HOME]} is the same as os.environ["HOME"] or $HOME in the shell.
  • {this.current_env} is the same as settings.current_env
  • {env[PROGRAM_NAME]} is the same as os.environ["PROGRAM_NAME"] or $PROGRAM_NAME in the shell.
  • {this.DB_NAME} is the same as settins.DB_NAME or settings["DB_NAME"]

so in your program

from dynaconf import settings

settings.DB_PATH == '~/DEVELOPMENT/calculator/mydb.db'

@jinja token

If jinja2 package is installed then dynaconf will also allow the use jinja to render string values.


export PROGRAM_NAME=calculator


DB_NAME = "mydb.db"

DB_PATH = "@jinja {{env.HOME}}/{{this.current_env | lower}}/{{env['PROGRAM_NAME']}}/{{this.DB_NAME}}"

so in your program

from dynaconf import settings

settings.DB_PATH == '~/development/calculator/mydb.db'

The main difference is that Jinja allows some Python expressions to be avaluated such as {% for, if, while %} and also supports calling methods and has lots of filters like | lower.

Jinja supports its built-in filters listed in Builtin Filters Page and Dynaconf includes aditional filters for os.path module: abspath. realpath, relpath, basename and dirname and usage is like: VALUE = "@jinja {{this.FOO | abspath}}"

Merging existing data structures

If your settings has existing variables of types list ot dict and you want to merge instead of override then the dynaconf_merge and dynaconf_merge_unique stanzas can mark that variable as a candidate for merging.

For dict value:

Your main settings file (e.g settings.toml) has an existing DATABASE dict setting on [default] env.

Now you want to contribute to the same DATABASE key by adding new keys, so you can use dynaconf_merge at the end of your dict:

In specific [envs]

database = {host="", user="default"}

database = {user="dev_user", dynaconf_merge=true}

database = {user="prod_user", dynaconf_merge=true}

also allowed the alternative short format

database = {host="", user="default"}

dynaconf_merge = {user="dev_user"}

dynaconf_merge = {user="prod_user"}

In an environment variable:

Using @merge mark

# Toml formatted envvar
export DYNACONF_DATABASE='@merge {password=1234}'

or @merge mark short format

# Toml formatted envvar
export DYNACONF_DATABASE='@merge password=1234'

It is also possible to use nested dunder traversal like:

export DYNACONF_DATABASE__password=1234
export DYNACONF_DATABASE__user=admin
export DYNACONF_DATABASE__ARGS__timeout=30
export DYNACONF_DATABASE__ARGS__retries=5

Each __ is parsed as a level traversing thought dict keys. read more in environment variables

So the above will result in

DATABASE = {'password': 1234, 'user': 'admin', 'ARGS': {'timeout': 30, 'retries': 5}}
IMPORTANT lower case keys are respected only on *nix systems, unfortunately Windows environment variables are case insensitive and Python reads it as all upper cases, that means that if you are running on Windows the dictionary can have only upper case keys.

You can also export a toml dictionary.

# Toml formatted envvar
export DYNACONF_DATABASE='{password=1234, dynaconf_merge=true}'

Or in an additional file (e.g settings.yaml, .secrets.yaml, etc) by using dynaconf_merge token:

    password: 1234
    dynaconf_merge: true


      password: 1234

The dynaconf_merge token will mark that object to be merged with existing values (of course dynaconf_merge key will not be added to the final settings it is just a mark)

The end result will be on [development] env:

settings.DATABASE == {'host': '', 'user': 'dev_user', 'password': 1234}

The same can be applied to lists:


plugins = ["core"]

plugins = ["debug_toolbar", "dynaconf_merge"]


plugins = ["core"]

dynaconf_merge = ["debug_toolbar"]

And in environment variable

using @merge token

export DYNACONF_PLUGINS='@merge ["ci_plugin"]'

or short version

export DYNACONF_PLUGINS='@merge ci_plugin'

comma separated values also supported:

export DYNACONF_PLUGINS='@merge ci_plugin,other_plugin'

or explicitly

export DYNACONF_PLUGINS='["ci_plugin", "dynaconf_merge"]'

Then the end result on [development] is:

settings.PLUGINS == ["ci_plugin", "debug_toolbar", "core"]

If your value is a dictionary:

export DYNACONF_DATA="@merge {foo='bar'}"

# or the short

export DYNACONF_DATA="@merge foo=bar"

Avoiding duplications on lists

The dynaconf_merge_unique is the token for when you want to avoid duplications in a list.


scripts = ['', '']

scripts = ['', '', '', 'dynaconf_merge_unique']
export DYNACONF_SCRIPTS='["", "", "dynaconf_merge_unique"]'

The end result for [development] will be:

settings.SCRIPTS == ['', '', '', '', '']
Note that is set 3 times but it is not repeated in the final settings.

Known caveats

The dynaconf_merge and @merge functionalities works only for the first level keys, it will not merge subdicts or nested lists (yet).

For deeper nested objects use dunder merge.

Global merge


or put it in your .env file then Dynaconf will automatically merge all existing variables.

BEWARE: Using MERGE_ENABLED_FOR_DYNACONF can lead to unexpected results because you do not have granular control of what is being merged or overwritten so the recommendation is to use other options.

More examples

Take a look at the example folder to see some examples of use with different file formats and features.