Skip to content
UltiVoice Docs

What is UltiVoice?

UltiVoice is a Windows desktop application that automates the full video dubbing pipeline using AI — from speech recognition through translation to voice synthesis — without sending your video files to any cloud service.

What it does

Section titled “What it does”

Given a video file, UltiVoice:

  1. Transcribes the original audio using Whisper large-v3 (runs locally).
  2. Translates the transcript into your target language(s) — 50+ languages supported.
  3. Synthesises natural-sounding speech in the target language, including optional voice cloning.
  4. Mixes the dubbed audio back with the original video and optionally burns in subtitles.

All processing happens on your machine. Your videos are never uploaded to a remote server.

Key features

Section titled “Key features”
FeatureDetails
Offline processingWhisper ASR, translation, and TTS all run locally. No video upload.
50+ languagesCovers major world languages including Vietnamese, English, Japanese, Korean, Spanish, French, German, and more.
Voice cloningClone a speaker’s voice from a short reference clip using F5-TTS-Vi technology.
Subtitle exportExport styled ASS/SRT subtitles or burn them directly into the output video.
Batch projectsOrganise multiple videos into projects; re-run individual segments without processing the whole file.
Local privacyNo account needed for processing. License validation is the only outbound network call during normal use.

Who is it for?

Section titled “Who is it for?”
  • Content creators who want to reach a multilingual audience without hiring a dubbing studio.
  • Localisation teams who need a fast first-pass dub for review before professional voice recording.
  • Businesses producing training or product videos for international markets.
  • Developers and researchers who need a local, scriptable dubbing pipeline.

How the pipeline works

Section titled “How the pipeline works”
Source video
    │
    ▼
[ASR — Whisper large-v3]  ← runs locally, no upload
    │  transcript + timestamps
    ▼
[Translation]  ← local Qwen2.5-7B or configurable engine
    │  translated segments
    ▼
[TTS / Voice Synthesis]  ← F5-TTS-Vi / OpenVoice v2
    │  dubbed audio track
    ▼
[Mux + Subtitle burn-in]  ← FFmpeg, local
    │
    ▼
Output video file (MP4 / MKV)

Next steps

Section titled “Next steps”