AI dubbing cost calculator

How AI dubbing cost stacks up

People assume dubbing is one line item. It isn't. To turn a video in English into the same video in Spanish, German and Japanese, three separate jobs run: transcription (speech → source text), translation (source text → target text) and text-to-speech (target text → spoken audio in the new language). Each leg has its own API and its own per-unit price, and the way they multiply is the whole story of what dubbing costs.

The bridge between "minutes of video" and "what you're billed" is characters. A natural speaking pace is roughly 150 words a minute, which is about 1,000 characters of text. So one minute of video generates ~1,000 characters that flow through translation and TTS. With reference rates of ~$10 per million characters for translation and ~$15 per million for TTS, that minute costs about $0.01 to translate and $0.015 to synthesise — plus a one-time $0.006 to transcribe.

Here's the key asymmetry: transcription runs once, no matter how many languages you target — you only need the source text a single time. But translation and TTS re-run for every language. Dub into one language and you pay 0.006 + 0.025 ≈ $0.03/min. Dub into five and transcription stays at 0.006 while translation+TTS multiply five-fold: 0.006 + 5×0.025 ≈ $0.13/min. That's why the "per extra language" number, not the headline rate, drives a multi-market localisation budget.

DIY pipeline vs managed tool — the trade-off. An all-in-one dubbing product (ElevenLabs Dubbing, HeyGen Translate, Rask) hides all three legs behind one upload and charges a flat rate — typically $0.30 to $1.00 per minute per language. That's roughly 10× to 30× the raw API cost. You're not overpaying blindly: the premium buys speaker diarisation, automatic timing, cloned or matched voices and, on the pricier tiers, lip-sync that moves the on-screen mouth to the new audio — none of which the cheap pipeline gives you for free. The rule of thumb: plain voiceover at volume favours the DIY stack; face-on-camera content that needs lip-sync and a consistent cloned voice favours a managed tool. Use the toggle above to see both for your exact video length and language count.

How to use it

1. Enter how many minutes of video you dub per month and how many target languages you need.
2. Pick build-your-own pipeline to price the raw APIs, or all-in-one tool and choose a representative per-minute rate.
3. Read the blended cost per minute and your monthly total at the top.
4. The breakdown table shows each leg (transcription, translation, TTS) and the comparison table puts DIY and managed side by side.

Common mistakes

Multiplying transcription per language. You transcribe the source once — only translation and TTS scale with languages. Forgetting the character conversion. Both translation and TTS bill per character, so estimate ~1,000 characters per minute of speech, not per minute directly. Assuming DIY equals managed quality. The cheap pipeline gives you audio, not lip-sync or cloned voices — budget engineering time or a managed tool if those matter. Ignoring free tiers and bundles. Whisper-class transcription and many TTS voices have free monthly allowances that cover small channels entirely.

FAQ

Estimate only. Dubbing-pipeline prices are reference figures and vary by provider, voice, language pair and plan — verify in each provider's pricing page.