Dub a 10-minute video into 12 languages: ElevenLabs vs STT+TTS chain

Dubbing a 10-minute video into 12 languages costs $60.00 with ElevenLabs Dubbing v1 at $0.50 per minute, or $264.00 with v2 at $2.20 per minute. A do-it-yourself chain on Sume (one transcription, 12 text-to-speech jobs, 12 Timeline renders) is $17.23, not counting the translation text, which you supply.

The dubbing prices are from the ElevenLabs API pricing page (read 2026-10-04): v1 is $0.33 per minute with a watermark or $0.50 without, and v2 is $2.20 per minute.

The Sume chain, line by line

Transcribe once: a 10-minute file is one speech-to-text job at $0.01 per minute, $0.10. Translate the transcript into 12 languages with whatever text model you use. Then, per language, synthesize about 9,000 characters (an assumed length for 10 minutes of speech) at $0.0475 per 1K, 9 x $0.0475 = $0.4275, so 12 languages is $5.13.

To put each voice track on the picture, render through Timeline 1.0 with the track as the audio spine and the video as the slot. A render is $0.10 per output minute rounded up, so 10 minutes is $1.00 per language, $12.00 for 12. The total is $0.10 + $5.13 + $12.00 = $17.23.

What you give up

This is voiceover replacement, not voice cloning with lip movement. ElevenLabs Dubbing detects speakers and preserves a voice across languages in one step; the chain above requires you to handle translation, timing and speaker choice yourself. If your 9,000 characters of translated text run longer than the original speech, the audio will overrun the picture and you will need to tighten the script.

The chain's cost is also low enough to leave room for review. If a native reviewer checks each language for 15 minutes, the review is likely the bigger line than the $1.44 of machine cost.

• Pick ElevenLabs when you need speaker-aware, one-click dubbing.
• Pick the chain when the source is a single narrator and you want control of the script.

Steps

• Transcribe the original audio and review it.
• Translate and shorten per language so the speech fits 10 minutes.
• Submit 12 TTS jobs, each under the 20,000-character limit.
• Plan each Timeline render with the unbilled plan endpoint, then render.

The marginal language

Adding a thirteenth language costs $0.4275 + $1.00 = $1.4275 on the chain, since the transcript is already done. The same language costs $5.00 on ElevenLabs v1 and $22.00 on v2. The chain's cost is dominated by the Timeline render ($1.00), so if you only need an audio track, skip the render and the language costs $0.4275.

Caveats

The 9,000-character figure is an assumption; speaking rates vary by language. Dubbing prices were read on 2026-10-04 and ElevenLabs may change its tiers. Translation model costs are not included.