Extension .csv
Smartling Identifier CSV
Example File Example CSV with common directives
Resources Comma-Separated Values (CSV) files RFC

CSV is a delimited plain text file that uses a comma to separate values. It is supported by almost all spreadsheets and database management programs, including Google Sheets, Apple Numbers, LibreOffice Calc, and Apache OpenOffice Calc.

Converting Files to CSV

It is important that the CSV is encoded in Unicode (or UTF-8) to preserve any special characters in your content. Most programs are encoded as Unicode or UTF-8 by default, but the most common spreadsheet program, Microsoft Excel, has some restrictions.

To create a CSV from Excel, simply save-as CSV UTF-8 from the file format dropdown.


For information on how to successfully open a CSV in Microsoft Excel, see How to open a CSV file safely in Excel.

Content Parsing


If you update a CSV file without any custom Smartling directives, Smartling will:

  1. Ingest each cell as a string for translation
    • The string order is in order of row number, regardless of column order. For example, if the CSV has content in cells A1, C2 and B3, the strings in Smartling will be displayed in that order.
    • If using a non-native program (such as Microsoft Excel), only the content on the first sheet is ingested, excluding the tab name
  2. Capture HTML and parses with basic HTML parsing
  3. Not capture formulas or formatting
  4. Not capture string key and variant data
  5. Not create placeholders in the strings using a default regular expression to capture a number of common placeholder formats.
  6. Not capture any string instructions
  7. Deliver translations, one file per language or all languages in one file.
    • For the latter, each translation is listed in the cells under the source string, with no language indicator.

For more information, see Content Parsing.


If a string contains HTML, Smartling automatically parses the strings in the file with basic HTML parsing, even without any HTML directive. Just note, this basic parsing will not break down the CSV value with HTML into smaller strings. It is ingested as one string.

If this is not desirable, and you want the value to be broken down into smaller strings in Smartling based on the HTML formatting, you can use the directive:

  • string_format_paths with the value of HTML:*.

This will benefit both your translation resources and your translation memory leverage.

If you have given each CSV string a key, it is worth noting that in doing so, the additional smaller strings will not have any "keys", however, you still have the same "variants" across each of the additional smaller strings. In most cases, such as translation Jobs, this is acceptable, however, this is not acceptable if you are importing translations from a file, as keys are critical for alignment.

Example CSV string with default vs HTML parsing:

<p>This <br>is</br> another example source string</p><h1>This is a header <i>maybe</i>?<p>One more for the <strong>road</strong></p>




(1 string with variant only. Keys must be applied before translations can be imported)



(5 strings with variant only - Keys must be applied before translations can be imported)

<p>This <br>is</br> another example source string</p><h1>This is a header <i>maybe</i>?<p>One more for the <strong>road</strong></p>
  1. <p>This <br>is</br> another example source string</p>
  2. <h1>This is a header <i>maybe</i>?
  3. <p>One more for the <strong>road</strong></p>


Smartling currently does not support plural strings in CSV files.

Preparing CSV for Translation

Directives are commands embedded in code that essentially direct Smartling to the various elements in your file. Examples include which column is your content for translation, where to find translator instructions, where any key or variants are, and command multi-lingual output, so all translations are on one file. See examples of these directives below.

When applying directives to files, it's always best to use a text editor to ensure the directives are written in plain text (without formatting). Examples of text editors include SublimeText, TextEdit, TextWrangler, and Notepad++.

Specifying Paths

Some directives require you to specify a path or set of paths to keys or strings in the file. A path in CSV files is simply a column number, such as 1 (column A), 2 (column B) etc.

When declaring a path for a key or string instruction the key or instruction will be applied to the next translatable string to the right, so you will need to organize your files so that keys and instructions are to the left of translatable strings in each row.


A key is a unique identifier for a string. By default, keys are not generated automatically in CSV files. To apply a key to a string, a custom key path must be set. Use any column before the source path column to define a key for each string.

Set using the source_key_paths directive.

The value is the column number, e.g.: 1 for column A or 2 for column B.

String Instructions

String instructions help linguists understand the content that they are translating by providing additional helpful information about the content and how it should be treated.

Set using string_instructions_paths directive.

You cannot set both string_instructions_paths and string_format_paths for HTML at the same time; if you want to use HTML parsing you will need to add instructions to strings via the Smartling dashboard. 

Placeholder Format

If you do not specify a custom placeholder format, even if you specify other directives, Smartling will not convert the following subtext of a string into a placeholder for that string:

  • {x}
  • {{x}}
  • ${x}
  • %x%
  • %%x%%
  • ##x##
  • __x__

To successfully convert any subtext of a string into a placeholder, the  placeholder_format_custom directive must be used. The value depends on the format of the subtext, or placeholder. For more information and example placeholder formats with matching values, see Placeholders in Resource Files.

Other Information

You may define values with and without quotations. For example:


value1, "Value 2"

If you want to use the symbol “ inside quoted value you escape it with double quotes like:


"She said ""hello"" to me."

This corresponds to the string: She said "hello" to me

For download options and how to open a CSV in Excel, see Translated CSV Files.


File directives are supported, both inline and via our APIDirectives are specified in comments within the files, in the following format:

Inline File Format

# smartling.[directive_name] = [value] or [path]

API Parameter

smartling.[directive_name] = [value] 

Here are some examples of [directive_name], along with example values or paths.


Values  column number - e.g.: 1

Applies a character limit to the translation which will be visible on the string details in the dashboard and the CAT Tool. The character limit column should be placed somewhere before the source column. 

The directive should point to the column that contains a character limit for the string.

Each string character limit is applied to the next source string, so you must place the character limit column to the left of (or before) the source string.

The character limit applies only to the first translatable column placed after the character limit field. You can apply a character limit for each string by inserting a character limit number alongside each string cell.

To remove a character limit from a CSV file that had been previously applied, the number in the character limit column must be set to "none". Simply removing the limit number from the column will not be successful.

Examples  # smartling.character_limit_paths = 1



Values Comma-separated list of columns.

Specifies which columns contain string instructions. This directive must be used together with smartling.paths to specify translatable strings.

Each string instruction is applied to the next translatable string, so you must place your instruction column to the left of (or before) the source string. You may have more than one instruction column per translatable string.

 Ignored if string_format_paths is set to HTML.

Smartling will capture the content in the files as follows. Column 1 will be captured as key metadata, Columns 2 and 3 will be string instructions. Column 4 contains the translatable strings.

# smartling.string_instructions_paths=2



Character limits and string instructions must precede source content.



Values  Values of all columns to be captured as strings. 
Description  Defines the column numbers with values to be captured as translatable strings. For multi-lingual translations import it defines a column and locale.


For uploading original file:
Comma-separated list of column numbers. # smartling.paths=1,2

For multi-language imports:
Comma-separated list of column/locale pairs # smartling.paths=1/[LocaleID],2/[LocaleID]


# smartling.paths=2,3

Specifies that columns 2 and 3 of the uploaded CSV file should be ingested as translatable strings.

# smartling.paths=2/es-ES,3/fr-FR

When importing translations, specifies that column 2 contains Spanish-SPAIN translations and column 3 contains French-FRANCE translations.




A comma separated list of paths to use create “keys” for strings on translate_paths.

The key will be a space separated string of all the keys leading to the source string. For example: “string”, “group1 string”.


Used to define the schema for capturing a key for each source string. Keys are required:

If you want to import pre-existing translations from a file with the same structure
If you want to create variants of strings that would otherwise be duplicates (By default Smartling does not create duplicate strings.)

Creating or updating variants for previously uploaded strings cause new strings to be created that will not have translations. The SmartMatch feature can be configured to automatically apply the existing translations, or translators can use the 100% match from the Translation to manually apply the translation.

Specify the full path to the value, then indicate which part of the path should be used as the key using {} notation.


# smartling.source_key_paths = 1

Smartling will capture data from column 1 as keys. Each key will be applied to the next translatable string after it, so keys need to be placed to the left of translatable strings in each row for this directive to work.




Description Used to specify a standard placeholder format.

# smartling.placeholder_format = IOS

Specifies iOS-style placeholders for the file.



Values 1) Custom Java regular expression.
2) NONE - disables any current custom placeholders
Description Specifies a custom placeholder format. Any text in your file matching the regular expression you provide will be captured as a placeholder.

# smartling.placeholder_format_custom = REGEX

# smartling.placeholder_format_custom=\{([^}]+)\}

Any characters surrounded by curly brackets, e.g., {first name}, will be treated as a placeholder.


See Placeholders in Resource Files for more on placeholders.



Values  true / TRUE or false / FALSE (default) 
Description  If TRUE, the first non-empty string in a CSV file will be treated as a header and excluded from translation. 
Examples  # smartling.first_row_header=TRUE 



Values integer - Accepted values are 0 - 100 


Sets the percentage by which original strings are inflated when downloading pseudo translations. If this directive is not set, pseudo translations are 30 percent longer than the original strings.’



# smartling.pseudo_inflation = 80

Downloaded pseudo translations will increase the length of original strings by 80 percent.




Values String of characters. The default is the comma ",".
"Description  Defines the sequence of characters that separate values in a record line. 

# smartling.field_separator=,

Fields are separated with a , character.



Values  String of characters. The default is the double quotation marks " 
Description  Defines the sequence of characters that may enclose values. To use the character sequence inside values you should escape it by repeating twice (default is ""). 



# smartling.string_encloser="

String literals are inclosed in " characters



Download Options



Values  true / TRUE or false / FALSE (default) 
Description  Defines whether all Smartling directives in the source file should be removed from translated files when downloaded. 
Examples  # smartling.strip_instructions_on_download=TRUE 



Values  Alternative labels for Smartling locales in JSON format. 
Description  Defines how languages are labeled in downloaded CSV files. The default label is the Smartling locale code, such as “fr-FR”, but you may wish to choose a different label, such as “French” in order to make the file easier to read or to match the labels used in your application. 

# smartling.locales_map={"es-ES":"Spanish","de-DE":"German"}

Downloaded translations will be labeled as Spanish for es-ES and German for de-DE.




Values  true / TRUE or false / FALSE (default) 

Determines whether to force the addition of a UTF-8 Byte Order Mark (BOM) to the output file when downloading translations.

If set to FALSE (default), output files will only include a BOM if the original file did.

If set to TRUE, a UTF-8 BOM will be added to the output file, even if none existed in the original file.

Note: This applies to UTF-8 only. For UTF-16, BOM is always used.

Examples  # smartling.add_utf8_bom=TRUE 



Values  true / TRUE or false / FALSE (default) 
Description  Defines if the original source strings should be included when downloading multiple languages.
Examples  # smartling.output_original_row=TRUE



Values  Column Number - for example: 4 


Used for “download multiple languages by row” option. Defines a column to record the language for each row. Output will display a language code for each column, e.g. de, en, es, etc.

This column should exist in the original file as an empty column.

If using this directive, you should also use smartling.paths to exclude the language path column from translation.



# smartling.translation_language_path = 4

When the translated file is downloaded, column 4 will record the language for each row.




Values  true / TRUE or false / FALSE (default) 
Description  Defines if the translated file should include all translations in one file, a locale per column. 

# smartling.translations_in_columns=TRUE

Need a template to start off with? Download this example source file. This template is designed to help you get started with a typical use case, but it's likely that you will need to adjust the file and add or remove directives to align with your need.


Translating CSV Files

Ensure to create a Files Project for file translation management.

Once you're ready to translate the file, create a Job. All content in the file will be ingested for translation.

By default, no Visual Context will be displayed to Translator from within the CAT Tool. If you would like to provide Translators with Visual Context, upload an image of the content's message. 

If you haven't provided string instructions from within the file using the directive above, you can also provide instructions from the dashboard to provide context.

By attaching a JPG or PDF document for reference, the Translators can download the attachment in the CAT Tool.

If you haven't applied character limits to the content from within the file using the directives above, applying character limits to strings in the dashboard can help done to ensure translations are kept to a certain length. 

When translations are complete, download the published translations to your locale drive.

Translated CSV Files

If you are using a program that is encoded with Unicode (or UTF-8) by default, then proceed to open the file as you would normally. In most cases, Microsoft Excel will be your default program for spreadsheet files. A simple double-click on your downloaded CSV could open your translated file in Excel with multiple corrupt characters. This is because Excel is not encoded with Unicode by default.

How to open a CSV file safely in Excel

After you have downloaded translations from Smartling

  1. Open a new blank workbook in Excel (separately)
  2. In Excel, go to the Data tab, click From Text. Choose the translated CSV file from your local drive and click Get DataScreenshot_2020-06-12_at_12.35.06.png
  3. Depending on your version of Excel, there is a series of steps to follow in the Import Wizard. Ensure Delimited is selected
  4. In the File Origin dropdown, scroll down and choose Unicode (UFT-8) > Next​​​​​​Screenshot_2020-06-12_at_12.48.56.png
  5. Ensure the Delimiters are set to Comma > Next​​Screenshot_2020-06-12_at_12.49.12.png
  6. Ensure the Column data format is Text > FinishScreenshot_2020-06-12_at_12.49.21.png
  7. ​Choose where you want the data > OKScreenshot_2020-06-12_at_12.49.31.png
  8. The CSV should have imported successfully. If you find corrupt characters in the file that are not visible in the Smartling Dashboard, revert to Microsoft Excel documentation

Was this article helpful?