Skip to content

Configuration File

The various Crowdin tools use a crowdin.yml configuration file that specifies the resources to be managed, including the files to be uploaded to Crowdin and the locations of the corresponding translations.

Configuration File Structure

A valid crowdin.yml configuration file has the following structure, so please ensure that you fill out all the needed information:

  • Head of the file with project credentials, preferences, and access information.
  • One files array element that describes a set of source and translation files you will manage.
  • Required fields in the files array: source that defines filters for the source files and translation with instructions on where to store the translated files or where to look for translations you already have if you want to upload them while setting up the CLI.

Writing a Simple Configuration File

A typical configuration file looks similar to the following:

crowdin.yml
"project_id": "projectId"
"api_token": "personal-access-token"
"base_path": "."
"base_url": "https://api.crowdin.com"
"preserve_hierarchy": true
"files": [
{
"source": "/locale/en/folder1/[0-2].txt",
"translation": "/locale/%two_letters_code%/folder1/%original_file_name%"
},
]
Name
Description
project_idCrowdin Project numeric ID
api_tokenCrowdin Personal Access Token. The token owner should have at least Manager permissions in the project
base_path
optional
Path to your project directory on a local machine relative to the crowdin.yml configuration file
base_url
optional
Crowdin API base URL. Optional for crowdin.com. For Crowdin Enterprise use the https://{organization-name}.api.crowdin.com
preserve_hierarchy
optional
If set to true, the directory structure will be preserved on the server. If set to false, the directory structure will be flattened.
sourceFilter for the source files. Wildcards are supported.
translationPath to store the translated files. Wildcards are supported.

API Credentials from Environment Variables

You could load the API Credentials from an environment variable, for example:

crowdin.yml
"project_id_env": "CROWDIN_PROJECT_ID"
"api_token_env": "CROWDIN_PERSONAL_TOKEN"
"base_path_env": "CROWDIN_BASE_PATH"
"base_url_env": "CROWDIN_BASE_URL"

If mixed, api_token and project_id are prioritized:

crowdin.yml
"project_id_env": "CROWDIN_PROJECT_ID" # Low priority
"api_token_env": "CROWDIN_PERSONAL_TOKEN" # Low priority
"base_path_env": "CROWDIN_BASE_PATH" # Low priority
"base_url_env": "CROWDIN_BASE_PATH" # Low priority
"project_id": "projectId" # High priority
"api_token": "personal-access-token" # High priority
"base_path": "." # High priority
"base_url": "https://api.crowdin.com" # High priority

General Configuration

The sample configuration provided above has source and translation attributes containing standard wildcards (also known as globbing patterns) to make it easier to work with multiple files.

Here are patterns you can use:


* (asterisk)

Matches any character in the file or directory name. If you specify *.json, it will include all files like messages.json, about_us.json, and anything that ends with .json.


** (doubled asterisk)

Matches any string recursively (including subdirectories). Note that you can use ** in both the source and translation patterns. When you use ** in the translation pattern, it always includes a sub-path from the source for a given file. For example, you can use source: /en/**/*.po to recursively upload all *.po files to Crowdin Enterprise. The translation pattern will be /%two_letters_code%/**/%original_file_name%.


? (question mark)

Matches any single character.


[set]

Matchesany single character in a set. Behaves exactly like character sets in Regexp, including set negation ([^a-z]).


\ (backslash)

Escapes the next metacharacter.

Placeholders

Crowdin CLI allows to use the following placeholders to put appropriate variables into the resulting file name:

NameDescription
%original_file_name%Original file name
%original_path%Take parent folder names in the Crowdin Enterprise project to build file path in the resulting bundle
%file_extension%Original file extension
%file_name%File name without extension
%language%Language name (e.g., Ukrainian)
%two_letters_code%Language code ISO 639-1 (e.g., uk)
%three_letters_code%Language code ISO 639-2/T (e.g., ukr)
%locale%Locale (e.g., uk-UA)
%locale_with_underscore%Locale (e.g., uk_UA)
%android_code%Android Locale identifier used to name “values-” directories
%osx_code%OS X Locale identifier used to name “.lproj” directories
%osx_locale%OS X locale used to name translation resources (e.g., uk, zh-Hans, zh_HK)

You can also define the path for files in the resulting archive by putting a slash (/) at the beginning of the pattern.

Usage of Wildcards

Structure of files and directories on the local machine:

  • Directorybase_path
    • Directoryfolder
      • 1.xml
      • 1.txt
      • 123.txt
      • 123_test.txt
      • a.txt
      • a1.txt
      • crowdin?test.txt
      • crowdin_test.txt
    • 1.xml
    • 1.txt
    • 123.txt
    • 123_test.txt
    • a.txt
    • a1.txt
    • crowdin?test.txt
    • crowdin_test.txt
    • 3.txt

Example 1. Usage of wildcards in the source path:

crowdin.yml
"files": [
{
"source": "/**/?[0-9].txt", # upload a1.txt, folder/a1.txt
"translation": "/**/%two_letters_code%_%original_file_name%"
},
{
"source": "/**/*\\?*.txt", # upload crowdin?test.txt, folder/crowdin?test.txt
"translation": "/**/%two_letters_code%_%original_file_name%"
},
{
"source": "/**/[^0-2].txt", # upload 3.txt, folder/3.txt, a.txt, folder/a.txt (ignore 1.txt, folder/1.txt)
"translation": "/**/%two_letters_code%_%original_file_name%"
}
]

Example 2. Usage of wildcards for ignoring files:

crowdin.yml
"files": [
{
"source": "/**/*.*", #upload all files that the base_path contains
"translation": "/**/%two_letters_code%_%original_file_name%",
"ignore": [
"/**/%two_letters_code%_%original_file_name%", #ignore the translated files
"/**/?.txt", #ignore 1.txt, a.txt, folder/1.txt, folder/a.txt
"/**/[0-9].txt", #ignore 1.txt, folder/1.txt
"/**/*\\?*.txt", #ignore crowdin?test.txt, folder/crowdin?test.txt
"/**/[0-9][0-9][0-9].txt", #ignore 123.txt , folder/123.txt
"/**/[0-9]*_*.txt" #ignore 123_test.txt, folder/123_test.txt
]
}
]

Renaming Translations File

If you need to rename a file with translations after the export, you can easily do this with the help of the parameter translation_replace.

For example, if the file is named strings_en.po, it can be renamed to strings.po. For this, add a new parameter (at the same level as the translation parameter) to the configuration file:

crowdin.yml
"files": [
{
"source": "/locale/**/*_en.po",
"translation": "/locale/**/%original_file_name%_%two_letters_code%",
"translation_replace": {
"_en": ""
}
}
]

In this case, _en will be erased from the file name.

Ignoring Files and Directories

From time to time, there are files and directories you don’t need to translate in Crowdin. In such cases, local per-file rules can be added to the config file in your project.

crowdin.yml
"files": [
{
"source": "/**/*.properties",
"translation": "/**/%file_name%_%two_letters_code%.%file_extension%",
"ignore": [
"/test/file.properties",
"/example.properties"
]
},
{
"source": "/locale/en/**/*.po",
"translation": "/locale/%two_letters_code%/**/%original_file_name%",
"ignore": [
"/locale/en/templates",
"/locale/en/workflow"
]
}
]

You can also use wildcards to ignore files.

Excluding Target Languages for Source Files

By default, the source files are available for translation into all target languages of the project. If some source files shouldn’t be translated into specific target languages, you can exclude them with the help of the parameter excluded_target_languages.

Configuration file example:

crowdin.yml
"files": [
{
"source": "/resources/en/*.json",
"translation": "/resources/%two_letters_code%/%original_file_name%",
"excluded_target_languages": [
"uk",
"fr"
]
}
]

Multilingual Spreadsheet

If a CSV or XLS/XLSX file contains the translations for all target languages, you should specify appropriate language codes in the scheme.

CSV file example:

identifier,source_phrase,context,Ukrainian,Russian,French
ident1,Source 1,Context 1,,,
ident2,Source 2,Context 2,,,
ident3,Source 3,Context 3,,,

Configuration file example:

crowdin.yml
"files": [
{
"source": "multicolumn.csv",
"translation": "multicolumn.csv",
"first_line_contains_header": true,
"scheme": "identifier,source_phrase,context,uk,ru,fr"
}
]

If your CSV or XLS/XLSX file contains columns that should be skipped on import, use none for such columns in the scheme, for example:

crowdin.yml
"scheme" : "identifier,source_phrase,context,uk,none,ru,none,fr"
Scheme Constants

To form the scheme for your CSV or XLS/XLSX file, use the following constants:

  • identifier – Column contains string identifiers.
  • source_phrase – Column contains source strings.
  • source_or_translation – Column contains source strings, but the same column will be filled with translations when the file is exported. When uploading existing translations, the values from this column will be used as translations.
  • translation – Column contains translations.
  • context – Column contains comments or context information for the source strings.
  • max_length – Column contains max.length limit values for the translations of the strings.
  • labels – Column contains labels for the source strings.
  • none – Column that will be skipped on import.

Saving Directory Structure on Server

You can use the preserve_hierarchy option to preserve or flatten the directory structure of your source files in the Crowdin project.

Example of the configuration file using the preserve_hierarchy option:

crowdin.yml
"preserve_hierarchy": true
"files": [
{
"source": "/locale/en/**/*.po",
"translation": "/locale/%two_letters_code%/**/%original_file_name%"
}
]

Let’s say the file/folder structure on your machine looks like this:

  • Directorylocale
    • Directoryen
      • Directoryemails/
      • Directoryapp
        • foo.po
        • bar.po

If you don’t use the "preserve_hierarchy": true option in your configuration file at all or use it with the false value, all shared directories will be skipped, and the file structure in Crowdin will be represented as follows:

  • Directory(root)/
    • foo.po
    • bar.po

Using the "preserve_hierarchy": true option, the file structure in Crowdin will be represented as follows:

  • Directorylocale
    • Directoryen
      • Directoryapp
        • foo.po
        • bar.po

The directories that don’t contain any files for translation won’t be created in Crowdin (i.e., as the emails directory in the example above).

Uploading Files to Specified Path with Specified Type

This feature adds support for two optional parameters in the yml file section: dest and type. It’s typically used for projects where the uploaded name must be different so that Crowdin can detect the type correctly.

The dest parameter allows you to specify a file name in Crowdin. It works for multiple files at once and supports the following placeholders: %original_file_name%, %original_path%, %file_extension%, %file_name%.

Example of the configuration file with both parameters:

crowdin.yml
"files": [
{
"source": "/conf/**/*.txt",
"dest": "/conf/**/%file_name%.properties",
"translation": "/conf/**/%two_letters_code%/%file_name%.properties",
"type": "properties"
},
{
"source": "/app/*.txt",
"dest": "/app/%file_name%.xml",
"translation": "/res/values-%android_code%/%original_file_name%",
"type": "android"
}
]

Changed Strings Update

You can use the update_option parameter to preserve translations for changed strings during the file update. If it is not set, translations for changed strings will be lost. Useful for typo fixes and minor changes in source strings.

Depending on the value, update_option is used to preserve translations and preserve/remove validations of changed strings during file update.

Acceptable values:

  • update_as_unapproved - preserve translations of changed strings and remove validations of those translations if they exist
  • update_without_changes - preserve translations and validations of changed strings

Example of the configuration with the update_option parameter:

crowdin.yml
"files": [
{
"source": "/*.csv",
"translation": "/%three_letters_code%/%file_name%.csv",
"first_line_contains_header": true,
"scheme": "identifier,source_phrase,translation,context",
"update_option": "update_as_unapproved"
},
{
"source": "/**/*.xlsx",
"translation": "/%three_letters_code%/folder/%file_name%.xlsx",
"update_option": "update_without_changes"
}
]

Custom Segmentation

Upload your XML, HTML, MD, or any other source files without a key-value structure with your own segmentation rules. If not specified, the pre-defined segmentation rules (SRX 2.0) are used for automatic content segmentation.

Example of the configuration file custom segmentation:

crowdin.yml
"files": [
{
"source": "/emails/sample1.html",
"translation": "/emails/%locale%/%original_file_name%",
"custom_segmentation": "/rules/sample.srx.xml"
}
]

Import Options

You can use additional parameters to customize the import process for specific file types.

XML Files Import Options

OptionTypeDescription
translate_content
optional
boolDefines whether to translate texts placed inside the tags.
Acceptable values are 0 or 1. Default is 1.
translate_attributes
optional
boolDefines whether to translate tags’ attributes.
Acceptable values are 0 or 1. Default is 1.
content_segmentation
optional
bool

Defines whether to split long texts into smaller text segments.
Acceptable values are 0 or 1. Default is 1.

Note! When Content segmentation is enabled, the translation upload is handled by an experimental machine learning technology.

translatable_elements
optional
array

This is an array of strings, where each item is the XPaths to DOM element that should be imported.
Sample path: /path/to/node or /path/to/attribute[@attr]

Note! If defined, the parameters translate_content and translate_attributes are not taken into account while importing.

Example of the configuration with additional parameters:

crowdin.yml
"files": [
{
"source": "/app/sample1.xml",
"translation": "/app/%locale%/%original_file_name%",
"translate_attributes": 1,
"translate_content": 0
},
{
"source": "/app/sample2.xml",
"translation": "/app/%locale%/%original_file_name%",
"translatable_elements": [
"/content/text", # translatable texts are stored in 'text' nodes of parent node 'content'
"/content/text[@value]" # translatable texts are stored in 'value' attribute of 'text' nodes
]
}
]

Other Files Import Options

OptionTypeDescription
content_segmentation
optional
bool

Defines whether to split long texts into smaller text segments. Only for TXT, DOCX, DITA, IDML, MEDIAWIKI, HTML, Front Matter HTML, Markdown, Front Matter Markdown, XML, XLIFF, XLIFF 2.0
Acceptable values are 0 or 1. Default is 1.

Note: When Content segmentation is enabled, the translation upload is handled by an experimental machine learning technology.

Example of the configuration with additional parameters:

crowdin.yml
"files": [
{
"source": "/emails/sample1.html",
"translation": "/emails/%locale%/%original_file_name%",
"content_segmentation": 1
}
]

Export Options

You can use additional parameters to customize the export process for specific file types.

Java .properties File Format

Escape Quotes

Defines whether a single quote should be escaped by another single quote or backslash in exported translations. You can add the escape_quotes per-file option.

Acceptable values:

  • 0 - do not escape single quote
  • 1 - escape single quote with another single quote
  • 2 - escape single quote with a backslash
  • 3 - escape single quote with another single quote only in strings containing variables (default)
Escape Special Characters

Defines whether any special characters (=, :, ! and #) should be escaped by a backslash in exported translations. You can add the escape_special_characters per-file option.

Acceptable values:

  • 0 - do not escape special characters
  • 1 - escape special characters by a backslash (default)

Example of the configuration:

crowdin.yml
"files": [
{
"source": "/en/strings.properties",
"translation": "/%two_letters_code%/%original_file_name%",
"escape_quotes": 1,
"escape_special_characters": 0
}
]

Configuration File for VCS Integrations

VCS integrations require the same configuration file as the CLI tool, meaning the same structure is supported. The only difference is that you should not store the project credentials in the file header for security reasons. Also, you can use a few additional parameters.

Pull Request Title and Labels

The default pull request title is New Crowdin updates. To specify your custom pull request title and add labels to the pull request, you can use the following parameters in the configuration file: pull_request_title, pull_request_labels.

crowdin.yml
"pull_request_title": "Custom PR title"
"pull_request_labels": [
"crowdin",
"l10n"
]

Commit Message

Each time translations are committed the default message is shown New translations {fileName} ({languageName}). You can use the commit_message parameter to add Git tags (e.g., to skip builds).

crowdin.yml
"commit_message": "[ci skip]"

To replace the default commit message, use the append_commit_message parameter with the false value. You can also add two optional placeholders: %original_file_name% and %language% to use the appropriate file name and language variables accordingly.

crowdin.yml
"commit_message": "Fix: New translations %original_file_name% from Crowdin"
"append_commit_message": false

Export Languages

By default, all the languages are exported. If you need to export some specific languages, use the export_languages parameter to specify them.

crowdin.yml
"export_languages": [
"uk",
"ja"
]

Pull Request Assignee

If you need to assign a pull request to particular users, use the pull_request_assignees parameter to specify them.

GitHub/GitHub Server:

crowdin.yml
"pull_request_assignees": [
"login1",
"login2"
]

GitLab/GitLab Self-Managed:

crowdin.yml
"base_path": "."
"pull_request_assignees": [
"user_id1", # numeric ID
"user_id2" # numeric ID
]

Pull Request Reviewer

If you need to request a pull request review from particular users, use the pull_request_reviewers parameter to specify them.

GitHub/GitHub Server:

crowdin.yml
"pull_request_reviewers": [
"login1",
"login2"
]

GitLab/GitLab Self-Managed:

crowdin.yml
"pull_request_reviewers": [
"user_id1", # numeric ID
"user_id2" # numeric ID
]

Bitbucket:

crowdin.yml
"pull_request_reviewers": [
"uuid1", # user uuid
"uuid2" # user uuid
]

Bitbucket Server:

crowdin.yml
"pull_request_reviewers": [
"username1",
"username2"
]

Azure Repos:

crowdin.yml
"pull_request_reviewers": [
"guid1", # user ID
"guid2" # user ID
]

Adding Labels to Source Strings

To add existing or new labels to the source strings, use the labels parameter. Labels will be added to the source strings only during the initial upload to the Crowdin project.

The strings uploaded to the Crowdin project before the use of the labels parameter won’t be labeled. If you remove the label added during the initial upload directly in Crowdin, it won’t be re-added on the next syncs.

Example:

crowdin.yml
"files": [
{
"source" : "/resources/en/*.json",
"translation" : "/resources/%two_letters_code%/%original_file_name%",
"labels" : [
"android",
"emails"
]
}
]

Read more about Labels.

Language Mapping

Often software projects have custom names for locale directories. Crowdin allows you to map your own languages to be recognizable in your projects.

Let’s say your locale directories are named en, uk, fr, de. All of them can be represented by the %two_letters_code% placeholder. Still, you have one directory named zh_CH. You can also override language codes for other placeholders like %android_code%, %locale%, etc.

To make it work with Crowdin without changes in your project, you can set up Language Mapping via UI.

Using One Configuration File for VCS Integrations and CLI

There are cases when it’s necessary to use VCS integration and CLI for one project. Mostly, in this kind of situation, you’d need to have two separate configuration files, one for VCS integration and another for CLI. However, you can use a single configuration file for both cases.

Since the VCS integration configuration file doesn’t contain project_id and api_token credentials required for CLI, you can pass them directly in the command using the following parameters: -i/--project-id, -T/--token.

As a result, your command for downloading translations via CLI will look like the following:

Terminal window
crowdin download -i {your-project-id} -T {your-token}

Example Configurations

Uploading CSV files
crowdin.yml
"project_id": "projectId"
"api_token": "personal-access-token"
"base_path": "."
"base_url": "https://api.crowdin.com"
"files": [
{
"source": "/*.csv",
"translation": "/%two_letters_code%/%original_file_name%",
# Specifies whether first line should be imported or it contains columns headers
"first_line_contains_header": true,
# Used only when uploading CSV file to define data columns mapping
"scheme": "identifier,source_phrase,translation,context,max_length"
}
]
GetText Project
crowdin.yml
"project_id": "projectId"
"api_token": "personal-access-token"
"base_path": "."
"base_url": "https://api.crowdin.com"
"files" : [
{
"source" : "/locale/en/**/*.po",
"translation" : "/locale/%two_letters_code%/LC_MESSAGES/%original_file_name%",
"languages_mapping" : {
"two_letters_code" : {
"zh-CN" : "zh_CH",
"fr-QC": "fr"
}
}
}
]
Android Project
crowdin.yml
"project_id": "projectId"
"api_token": "personal-access-token"
"base_path": "."
"base_url": "https://api.crowdin.com"
"files" : [
{
"source" : "/res/values/*.xml",
"translation" : "/res/values-%android_code%/%original_file_name%",
"languages_mapping" : {
"android_code" : {
"de" : "de"
}
}
}
]
Was this page helpful?