Dubbing
Translate spoken audio — or a whole video — into another language while preserving each speaker’s voice and identity.
Overview
The Dubbing node uses ElevenLabs Dubbing to translate media from one language to another, preserving each speaker’s voice characteristics. It is dual-mode:
- Audio in → dubbed audio out (the classic mode).
- Video in → dubbed VIDEO out, plus the dubbed audio track on the audio output handle. The dubbed clip keeps the original visuals with the translated dialogue.
- Source link — instead of wiring an input, paste a public link (YouTube, TikTok, or a direct media URL) in the panel. ElevenLabs fetches the link itself; the file never passes through Nodaro’s servers. The result mode (audio/video) follows what the link points at.
The process is asynchronous — the node submits the dubbing job, polls for completion, and returns the translated media when ready. Long sources are delivered by the platform’s background recovery lane; the job completes even if it outlives the initial polling window.
Limits & Pricing
| Maximum dubbed span | 30 minutes — for longer sources, set a Start/End window to dub part of the file |
| Maximum uploaded file size | 500 MB (source links are exempt — ElevenLabs fetches those directly) |
| Price | 40 credits per minute of the dubbed span, minimum 1 minute |
The dubbed span is the Start/End window when set, otherwise the whole source. When the duration cannot be determined up front (source links, probe failures), a 2-minute reserve (80 credits) is held and the real duration is verified against the 30-minute cap once ElevenLabs probes the media.
Worked examples (matching the platform’s pricing tests):
- A 60-second clip → 1 minute → 40 credits.
- A typical 2-minute clip → 80 credits (exactly the old flat price).
- A 10-minute video → 400 credits.
- A 45-minute video with a window of 0:00–10:00 → the window is the span → 400 credits (without the window the request is rejected at 30+ minutes).
Hebrew dubbing
Hebrew (he, also accepted as heb) uses ElevenLabs’ project API with Dubbing v2, priced separately at 1,100 credits per started minute before any account-specific adjustments. Upload/import the source first and pass it as audioUrl or videoUrl; provider-fetched sourceUrl links are not supported for this target. The source video is retained and combined with the translated soundtrack so the result can be previewed and compared normally.
This target preserves voices and background audio and detects speakers automatically. Native voice substitution, a fixed speaker count, background removal, Start/End windows, provider watermark discounts, profanity filtering and target-accent overrides are rejected before dispatch. Trim the source before dubbing. Video keeps its source resolution. Existing jobs and other target languages continue using the legacy API.
Configuration
| Field | Type | Default | Description |
|---|---|---|---|
| Target Language | string |
"es" |
Language code for the desired output language (required) |
| Source Language | string (optional) |
auto-detect | Language code of the input audio. Leave empty for automatic detection |
| Source Link | string (optional) |
— | Public YouTube/TikTok/direct URL — ElevenLabs fetches it directly. Overrides any wired input |
| Start / End (sec) | number (optional) |
whole source | Dub only this window of the source |
| Number of Speakers | number (0-20, optional) |
0 = auto | Number of distinct speakers in the input; 0 or empty = auto-detect |
| Native voice | boolean |
false |
By default the dub clones the original speaker — they speak the target language with their own voice and accent. Enable to use a similar native-sounding Voice Library voice instead (clean target-language accent) |
| Drop background audio | boolean |
false |
Remove background audio from the final dub — improves quality for speech-only sources (speeches, monologues, voiceovers) |
| Keep source resolution | boolean |
false |
Render video dubs at the source’s original resolution (slower) |
| Profanity filter | boolean |
false |
Apply ElevenLabs’ profanity filter to the dubbed speech |
| Target Accent | string (optional) |
— | Experimental: steer the dubbed voices toward an accent |
Voice cloning vs native voice
The default mode preserves speaker identity: a Hebrew speaker dubbed to English sounds like the same person speaking English, including their accent. If you want the dub to sound like a native target-language speaker instead, enable Native voice — ElevenLabs then picks a similar voice from its Voice Library. Note: library voices used this way count toward the workspace’s custom-voice slots; if no slots are free the dub fails with an error.
Inputs & Outputs
- Inputs:
audio— source audio to dub;video— source video to dub (video wins when both are wired; a Source Link overrides both) - Outputs:
audio— the dubbed audio track (URL; always produced);video— the dubbed video (URL; video mode only)
The mode is decided by the media, not the slot: an audio-only file wired into the video input is dubbed as audio; a video wired into the audio input is treated as a request for audio.
Best Practices
- Explicitly set the source language when you know it — auto-detection is reliable but specifying it avoids edge cases with accented or mixed-language speech.
- Specify the number of speakers if the input has multiple voices. Auto-detection works but can occasionally merge or split speakers incorrectly.
- Use clean source audio without heavy background music. The model handles moderate background noise, but music overlapping speech degrades quality.
- For sources longer than 30 minutes, dub in windows (Start/End) — each window is its own run and its own per-minute charge.
- Test with shorter clips (or a short window) first before dubbing long media to verify the target language quality meets expectations.
Common Use Cases
- Localizing podcast episodes or full videos for international audiences
- Dubbing a YouTube/TikTok link into another language without downloading it first
- Creating multilingual versions of training or educational content
- Dubbing interview audio for cross-language content repurposing
- Translating voiceover tracks for marketing videos