Kotoba demos real-time voice cloning translation: Japanese to English with original voice
Tim_Dettmers · x · 2026-08-25
Kotoba Tech demonstrated a new feature in its real-time speech-to-speech translation system: on-the-fly voice cloning. The model picks up the speaker's voice and carries it into the target language (e.g., Japanese to English), preserving the original timbre. It is currently in Alpha stage.
More from Multimodal
- Jazzy vinyls: Generate jazz-style record covers and music — joonasvirtanen · 2026-08-25
- Flova agent workflow automates full commercial video production — egeberkina · 2026-08-25
- AI Art Short: The Winter Oath — Realistic-Stomach-97 · 2026-08-25
- Developer Creates Cyberpunk FPS Game Using Grok Build — tetsuoai · 2026-08-25
- The Witch Trial — a medieval AI short film — johnstro12 · 2026-08-25
- Testing LTX video reference with H3 model for motion transfer — cocktailpeanut · 2026-08-25