FAQ

Frequently Asked Questions


Key Management


Q: How do I create a new translation key?


R: Go to the key tree, click the "Create file or folder" button or use the context menu (right-click).


Enter the full path of the key and confirm.


The key will be created in the selected namespace.



Q: What is the difference between a folder and a file in the tree?


R: Folders are logical containers for organizing your keys.


Files are individual keys containing the text to be translated.


Folders can contain other folders or keys.



Q: How do I delete an unused key?


R: Activate the "Show/Hide orphaned keys" option in the tree header to identify keys present in

LSDE

but absent from your source code.


You can then delete them via the context menu.



Translation and LLM


Q: Which LLMs are available in

LSDE

?


R:

LSDE

supports 7 main providers: Anthropic, OpenAI, Mistral, Gemini, Deepseek, ElevenLabs, and a free service via temporary emails.


You can configure your API keys in the global settings,

Authentication section

.



Q: How can I use LLMs for free?


R:

LSDE

offers a free method for OpenAI by creating an account with temporary emails (temp-mail or 10minmail).


You will have approximately 70 tasks per hour with a limited quota.


You can also use Google services with Gemini, which offers free API keys.



Q: Are LLM translations reliable?


R: The results are relevant and satisfactory, but it remains essential to use a language professional for final verification.


LLMs excel at translation when provided with the appropriate context.



Q: How can I improve the quality of LLM translations?


R: Fill in your key's metadata with descriptions, variables, and contextual notes.


This information is integrated into LLM queries to ensure more relevant results.



Metadata and Context


Q: What are metadata used for?


R: Metadata provides additional context to translators, writers, and LLMs.


They include descriptions, allowed variables, user notes, and images to visually contextualize the content.



Q: How do I add images to keys?


R: Open the metadata window for the selected key and use the "Images" section.


Images are compressed and integrated into the .lsde file for easy sharing.



Q: Do LLMs have access to user notes?


R: No, user notes are intended only for composers and translators.


They are not transmitted to LLMs.



Code Scanner


Q: How do I configure the code scanner?


R: Go to the Patterns section (project settings).


Create a pattern with a regular expression (Regex) containing at least 1 capture group to identify your i18n keys in your codebase.



Q: Which programming languages are supported?


R: No intrinsic limitation.


You simply need to create the appropriate Regex to capture your keys, regardless of the language.


Use a tool like regexr.com to test your patterns.



Q: How do I identify missing keys in my code?


R: Activate the "Missing key" mode in the code scanner.

LSDE

will only display keys present in your source code but absent from the

LSDE

tree.


You can create them in batches.



Rendering and Display


Q: What is post-processing text rendering?


R: It's a system for customizing text display in real-time.


You can configure Regex patterns to capture text groups and assign them decorators (colors, icons, widgets) to improve readability.



Q: How do I add "widgets" to text?


R: Create a pattern with a Regex in the project settings.


Assign a capture group to a "Widget" type decorator.


The widget will only appear in active windows to optimize performance.



Q: How do I link variables to rendering?


R: In the project settings, Variables section, create a category and variables with "tags".


Configure your Regex patterns to capture these tags, and

LSDE

will automatically associate the decorators with the found instances.



Voice and Narration


Q: How do I generate voices for my dialogues?


R: You must first obtain an ElevenLabs API key (free plan available).


Configure your voice profiles in the project settings, then use the voice manager to generate narrations.



Q: What is the difference between SAD and MAD?


R:

SAD

(Single Actor Dialog) = one speaker per key.

MAD

(Multi Actor Dialog) = multiple speakers within the same key, identified by Regex tags.


MAD significantly reduces the number of keys to manage.



Q: How do I manage multiple characters in a dialogue?


R: In MAD mode, use tags like

{character} text

in your key.


Configure a Regex to capture the character ID and their text, then associate voice profiles with each character in the variables.



Q: What if I modified the text after voice generation?


R:

LSDE

detects changes and prompts you to revalidate the voices.


In MAD mode, you can reassign existing voice instances via drag-and-drop if you have moved or added characters.



Configuration and Settings


Q: How do I change the font or text size in the editor?


R: Go to Global Settings > User Interface > Typography.


You can adjust the size, spacing, font, and line height.



Q: How do I change the

LSDE

interface language?


R: Go to Global Settings > Accessibility > Language Selection.


You can configure 2 interface languages and switch between them with [F1].



Q: How do I import an existing project?


R: Go to the Localization Manager and use the "Smart Import" option.


Structure your folders according to the

../folder/lang/namespace.json

format, and

LSDE

will import them automatically.



Q: How do I protect my keys from modifications?


R: In the main editor, check the "Protect" box to mark a text as final.


LLM tasks will automatically ignore it.



Collaboration and Team


Q: How do I share my

LSDE

project with a team?


R: The

.lsde

file contains the entire project.


Share it via Git, email, or a cloud service.

LSDE

will automatically synchronize changes upon opening.



Q: How do I track a collaborator's ongoing tasks?


R: Activate the "Show/Hide keys with ongoing tasks" option in the tree.


You will quickly see keys assigned to active tasks or awaiting validation.



Performance and Optimization


Q: Why does the editor slow down with multiple languages displayed?


R: By default,

LSDE

disables complex renderings (widgets) in inactive windows to optimize performance.


If this is not an issue, you can disable this optimization in the global settings.



Q: How can I improve the code scanner's speed?


R: Select a specific key in the tree rather than displaying all results.


This filters the results and speeds up the scanner.


Also, avoid overly complex Regex with conditionals.