Streamlining prompt translation in ComfyUI with native TranslateGemma support

If you’re tired of switching tabs to DeepL or Google Translate just to build your prompts, this is worth integrating into your pipeline.

Why it’s useful for Prompt Engineering:

• Automated Prompt Translation: This is the main draw. You can input complex scene descriptions or abstract concepts in your native language, and the node handles the translation directly within your workflow before passing it to CLIP/T5. Because TranslateGemma is specialized for this, the nuance retention is excellent compared to generic smaller LLMs.

• Multimodal Capabilities: It supports image-to-text translation natively. You can feed reference images directly into the node to extract or translate embedded text.

• Zero-VRAM Mode for Chinese: For Chinese users, it includes a chinese_conversion_only mode using OpenCC. It handles Traditional/Simplified conversion instantly without loading the model weights, saving your GPU memory for generation.

Installation: You can find it easily in ComfyUI Manager. Just search for: comfyui translategemma and install.

link for details: https://github.com/rookiestar28/ComfyUI-TranslateGemma