Feature Description#

This tool is suitable for multilingual translation tasks for files such as subtitles and text. It supports various scenarios including batch processing, cache optimization, and multi-language export to help you complete localization work efficiently.

Multi-Language Translation: Translate Once, Output Multiple Languages#

You can translate the same file into multiple languages simultaneously, which is ideal for multilingual subtitles or internationalization projects:

• Example: Translate English subtitles into Chinese, Japanese, German, French, etc.
• Currently supports 50+ languages, with more to be added in the future.

Translation Cache: Reduce Redundant Translations, Improve Efficiency#

To avoid redundant requests to the translation API, this tool introduces a local caching mechanism:

• Cache Rule: Each translated segment is stored using a unique key composed of: Source Text_Target Language_Source Language_Translation API_Model Settings.
• Cache Hit Condition: The cache is only used when all parameters match exactly.
• Purpose of Cache: Avoids re-translating content, reduces the number of API calls, and improves translation speed.

If you do not wish to use the cache, you can disable it in the settings by unchecking "Use Translation Cache," or you can manually clear it by clicking the "Clear Translation Cache" button in the API settings.

Remove Characters After Translation: Automatic Content Cleanup#

After translation is complete, you can automatically remove specified characters or symbols, ideal for batch cleaning of music symbols, markers, etc. in subtitles:

• How to Use: In the "Remove Characters After Translation" input field, enter the characters you want to remove, separated by spaces for multiple characters
• Example: Enter ♪ ♫ 🎵 to remove music symbols from subtitles
• Use Cases:
  • Remove music markers from subtitles (♪, ♫, 🎵, etc.)
  • Remove HTML tags (such as </i>, <b>)
  • Batch delete specific punctuation or symbols
• Processing Timing: Executed uniformly after all translation and format processing is complete to ensure translation quality is not affected
• Cross-Tool Sharing: This setting is shared across Subtitle Translation, Markdown Translation, and JSON Translation tools, and is saved in local storage

Translation Modes: Automatic Mode Selection#

The tool automatically switches translation modes based on your upload action, requiring no manual selection:

• Batch Mode (Default): Supports uploading multiple files simultaneously, suitable for large-scale translation tasks. Translated results are automatically downloaded and saved.
• Single-File Mode: When you upload only one file or paste text, the tool enters single-file mode, which is ideal for quick previews or testing translation results. Note: In single-file mode, each new upload will overwrite the previous one.

Settings Import/Export#

Supports one-click export/import of current settings, making it easy to move between devices or migrate deployments:

• Exported Content: Includes all configurations (including sensitive information like API keys).
• Importing: Will overwrite your current configuration. Please be sure to back up your settings before proceeding.

Detailed Translation API Parameters#

To balance context with API limitations, you can customize the following key parameters:

Chunk Translation Size#

For non-AI translation APIs, the tool automatically splits text into chunks before translation to handle long documents or subtitles. The Chunk Translation Size refers to the maximum character count for each chunk. Below are the maximum character limits for common APIs:

Translation Service	Max Characters per Request
DeepL API	128,000 characters
DeepLX Free	1,000 characters
Azure Translate	10,000 characters
Google Translate (Web)	5,000 characters
Google Cloud API	30,000 characters

⚠️ Note: The web version of Google Translate can break line breaks, so chunked translation is not used for it.

Delay Time (ms)#

This controls the interval between requests for chunked translations, helping to avoid frequent request failures:

• It is recommended to set a higher delay when on a poor network or using a free API.
• For example: When using the free Azure Translate tier, a setting of 5000ms or more is recommended.

Translation Rate / Context Window#

The "Translation Rate / Context Window" parameter serves two functions:

• Concurrent Request Limit: Controls the number of lines/requests sent to the API per batch. Higher values result in faster speeds but increase the likelihood of being rate-limited or triggering errors such as HTTP 429 or empty responses.
• Context Window Size: In the "Context-Aware Translation" mode for subtitles, this same parameter determines the number of preceding and succeeding lines included (subject to a safety cap). Setting this too high not only consumes more time but also makes it easier to exceed the LLM's context limit.

Tuning Suggestions:

• For standard APIs, it is recommended to adjust incrementally between 10 and 30; if errors occur frequently, prioritize lowering the rate or increasing the delay.
• If you need to increase the window in context mode, adjust it upwards slightly (e.g., 10%~15%), but avoid jumping to hundreds or thousands of lines at once to prevent the prompt from becoming excessively long.

Once stable, gradually increase the values while monitoring the API error rate and response time to ensure they remain within an acceptable range before relaxing the limits further.