AI FAQ

We do not recommend saving Generative Subtitles translations to your TMs. Subtitle content typically consists of short, fragmented segments that reflect speech patterns, pauses, and scene changes - making them unsuitable for reuse in traditional translation workflows.

 

For this reason we use large language models (LLMs) to translate subtitle content. LLMs excel at understanding context across fragmented input and producing fluent, subtitle-specific translations, outperforming traditional machine translation models in this area.

 

While TMs are used to provide contextual input to the LLM, the generated subtitles are optimized for readability and timing rather than for integration back into standard TM-based projects.