Video Translate: from recording to translated media

Quick start

1. 1

   Upload and choose languages

   Open Video Translate, select your video or audio, then choose the source and target languages. Add context or glossary terms for names and specialist vocabulary.

2. 2

   Choose the output and review

   Select Subtitles only, Dub only, or Subtitles + dub. Enable transcript review if needed, check the credit estimate, and start. If the job pauses for review, save corrections and approve the transcript to continue.

3. 3

   Preview, refine, and download

   Open the result from your job history. Check wording and timing, use the edit panel for corrections, and download the available SRT or translated video/audio for your selected output type.

Files, languages, and credits

Upload MP4, MOV, AVI, or MKV video, or MP3, M4A, or WAV audio. Check the upload size, duration limit, and credit information shown in your workspace before starting. Limits and available services depend on your account and configuration.

The language selectors offer 18 options: Simplified Chinese, Traditional Chinese, English, Spanish, French, German, Japanese, Korean, Portuguese, Russian, Arabic, Italian, Thai, Vietnamese, Indonesian, Turkish, Polish, and Dutch. Voice and recognition support can vary by provider.

Choose subtitles or dubbing

Subtitles only produces target-language SRT and skips speech synthesis. Dub only produces translated speech in video or audio. Subtitles + dub requests both. Audio uploads produce audio rather than a new video. Available voices depend on the selected language and service.

Original subtitles and SRT input

Choose Original subtitles only when you need an original-language transcript without translation or dubbing. This runs speech recognition and creates an SRT; open Details to review, copy, or download it.

Already have subtitles? Add an Optional SRT timeline before uploading the media. Choose Original-language SRT for text that still needs translation, or Already-translated SRT for text ready for speech. This skips recognition and opens the TTS editor for dubbing. Use a file no larger than 1 MB with at most 5,000 nonempty cues; resolve timeline validation errors before continuing. Leave this empty for the standard recognition workflow.

Review text, timing, and voices

Enable “Require human transcript review before translation” before starting if you want to correct recognition errors first. A job waiting for transcript review needs your approval before translation continues. Context and glossary hints help with terminology, but still check the result.

In the edit panel, review translated segments, speaker assignments, and timing. After changing dialogue, regenerate the affected speech and preview it before rendering. SRT imports show a matching preview: preserve current timing or adopt the uploaded cue structure, then review speaker mappings and uncovered lines. Version history lets you inspect and restore earlier edits.

Processing and provider choices

Classic mode uses the regular workflow. Agent mode adds recognition, translation, and speech quality checks with limited retries and a different credit rate. Choose only services shown as available. ElevenLabs Full Dubbing follows a separate provider workflow and skips the internal segment pipeline; the review and edit steps above describe the standard workflow. Full Dubbing does not currently accept Thai as a target; choose a supported segment TTS service for Thai instead.

Troubleshooting

• Upload rejected: check the file format, size, and duration against the limits displayed for your plan.
• Job waiting for review: open the transcript, save corrections, and approve it to resume processing.
• Speech sounds rushed: shorten the translated line or adjust its timing, then regenerate that segment and preview again.
• A provider is unavailable: choose an available service or contact your administrator. Let any automatic speech retries finish before retrying manually.