Translation that keeps context
The translator sees the whole conversation and the topic you give it, so tone and references carry through from one line to the next.
Upload a video once. Voixa transcribes it, translates the whole conversation in context and reads every line in a natural voice, timed to the scene. Fix any line you like, then export.
Real Voixa output: a 30-second English clip dubbed into Vietnamese in the voice Nhật Phong, with the translation polished in the Dub studio before export.
How it works
You upload a video and pick a voice. Voixa does the rest and shows each step as it runs.
Speech becomes text with a timestamp for every sentence. The source can be any language Whisper understands.
The whole conversation is translated together, not sentence by sentence, with your glossary and topic so names and terms stay right.
Each line is read in the voice you chose and placed at its moment on screen. Lines that run long are shortened or sped up slightly.
The new voice goes over the original audio, lowered to 10%, so music and room sound stay. Or keep the voice only.
Download an MP4 with subtitle tracks, plus SRT and VTT subtitles in both languages.
Features
The translator sees the whole conversation and the topic you give it, so tone and references carry through from one line to the next.
Every dubbed line starts where the original line starts. When a translation runs long, Voixa shortens it or speeds it up a little.
Filter the lines that need a look, rewrite a translation, and only that line is voiced again.
Save names, places and forms of address once per project, and every episode uses them.
A bilingual transcript follows the video line by line. Save words to a notebook and review them later.
Dub videos on YouTube, Udemy and other sites from their captions or audio, without downloading the video.
Quality control
In the Dub studio every line shows the original and the dub side by side, with a reason when it needs attention:
Use cases
Dub into:
From any language Whisper understands, including English, Japanese, Korean and Chinese.
Cost
Dubbing uses the same tokens as the rest of Voixa: a rate per minute of video plus a small amount per gigabyte uploaded. You always see the quote before the upload starts.
FAQ
Replacing the speech in a video with another language using AI. Voixa transcribes the speech, translates it, reads every line in a natural voice and mixes it over the original sound, then exports a finished video with subtitles.
It depends on the length of the video. The 30-second sample on this page took about three and a half minutes from upload to finished file, most of it starting the transcription machine. Long videos are transcribed in parts at the same time, so an hour of video does not take an hour.
Yes. The original audio stays under the new voice at 10% of its volume, so music, effects and room sound are still there. You can also export the new voice on its own.
Yes. Every line shows the original and the translation. Edit a line and only that line is read again. Lines that run too long for their scene are flagged.
Today the whole video is read in one voice that you choose, with every line placed where it was said. Pick a voice that suits the main speaker.
Voixa dubs into Vietnamese, English, French, German, Italian, Spanish, Portuguese, Chinese, Japanese and Korean. The source can be any language Whisper understands, including English, Japanese, Korean and Chinese.
MP4, MOV, WebM, MKV and M4V video, or M4A audio, up to 6 hours and 9 GB per file. You get an MP4 with subtitle tracks, plus SRT and VTT subtitles.
In tokens shared with the rest of Voixa: a rate per minute of video plus a small amount per gigabyte uploaded. You see the quote before the upload starts, and every account gets free tokens every day.
Yes, with the Voixa Chrome extension. It dubs from the video’s captions, or from its audio when there are none, and plays the dub alongside the original player.
Upload a video, check the quote, and listen to the result.