FAQ

Text translation models (Gemini, GPT, Claude, Grok, DeepSeek, Kimi, and more) detect text, translate it, and put the result back into the original layout. They cost 10 credits per image and are the default for comics, screenshots, documents, and product photos. Image generation models (Nano Banana 2 and GPT Image 2) redraw the translated text inside a new image for stronger visual matching. They cost 50 credits per image.

Start with Gemini 3.1 Flash Lite or another text model for most jobs — they are faster and use fewer credits. Switch to Nano Banana 2 or GPT Image 2 when stylized lettering, posters, or tightly designed graphics need the text painted back into the artwork.

Source language can be auto-detected. Target languages cover 100+ options, including English, Simplified Chinese, Traditional Chinese, Japanese, Korean, Spanish, French, German, and many more. Choose the pair in the translator before you start.

Yes. Text translation models let you pick a font such as NotoSans, WildWords, KomikaJam, or Bangers. Some fonts cover more writing systems than others, so NotoSans is the safest default for mixed languages.