Video dubbing method that adapts LTX 2.3 with a LoRA to translate speech and re-sync lips in a single model.
Updated EU status checked Last shown
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
JUST-DUB-IT is a research approach to video-to-video dubbing from the authors of the LTX audio-visual diffusion model line (Lightricks-affiliated researchers and Tel Aviv University collaborators). It adapts a pretrained joint audio-video diffusion model (LTX 2.3) with a lightweight LoRA. Given an input video, the model generates translated speech and matching facial motion together, instead of chaining separate translation, voice-cloning and lip-sync tools.
Training data is synthesized by the base model itself. It generates clips with a language switch partway through, then inpaints the face and audio in each half to match the other half's language, producing paired multilingual videos of the same speaker. Latent-aware masking limits motion leakage, and a structured cross-attention bias keeps audio and video signals from leaking into each other.
The project page compares results against LatentSync, X-Dub and HeyGen (the latter using ElevenLabs voice cloning) and reports better voice preservation, lip synchronization and naturalness. Examples cover French, Spanish and German. A YouTube review describes the LoRA as about 2.5 GB with code and model released, but the provided pages did not list repository or weights links, so those are unverified here. The paper is arXiv 2601.22143.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Video dubbing AI that translates speech into another language and adjusts lip movements. Built as a small 2.5 GB LoRA on LTX 2.3, reported to outperform methods like HeyGen. Code and model released.
Closest neighbours by meaning, not just by category.
Model
Expressive text-to-speech model with zero-shot voice cloning, derived from the LTX 2.3 audio model.
App · ElevenLabs
ElevenLabs' AI dubbing model that translates video and audio while carrying over the original speaker's voice and emotion.
Model
Research model for editing video and audio jointly from a text instruction, e.g. changing speech with lip sync or voice.
Model
Research model for real-time, audio-driven talking-head video generation from a reference image, using one-step inference.
Model · Tencent
Tencent Hunyuan's real-time Chinese-English speech-to-speech simultaneous translation model with voice cloning.
Library · Lightricks
Official Lightricks package for training and fine-tuning LTX-2 video models, including LoRAs for characters, effects and styles.
