Dub Your Talking-Head Reel Into Another Language With Lip-Sync: The Workflow and What It Costs (October 2026)

If you make talking-head Reels in one language, the cheapest growth available is the same video in another. You do not need to refilm: the voice can be translated and re-spoken in your own timbre, and the lips can be re-timed so the viewer does not see a mismatch. This guide walks through the workflow, the three ways to do it, what each costs for a typical 30-second Reel, and the limits that cause most failures.

What "dubbing" has to do

Three things happen, and a tool may do one, two or all three:

1. Translate the script into the target language.
2. Speak it in a voice that sounds like the original speaker.
3. Re-time the mouth so the lips match the new language.

A Reel where the speaker is on camera needs all three. A voice-over on b-roll needs only the first two.

Three routes

Prices read on 18 October 2026. For a 30-second talking head with the speaker on camera throughout, the one-call route is cheapest and simplest. The two-step route pays off on longer videos where the speaker is on screen for a minority of the time (the full comparison).

Step by step

1. Cut the source clean. One speaker, facing the camera, no hands across the mouth, minimal music under the voice. Background music complicates the audio separation.
2. Choose the language and, if you can, a glossary. Brand names and product terms should not be translated; HeyGen's glossary exists for this.
3. Run a short test. Ten seconds costs $0.50. Watch it with sound on a phone, and have a speaker of the language listen.
4. Check the length. A translation is often longer than the original. If a Spanish version needs 20 percent more time, the mouth will race unless the script is tightened.
5. Add captions in the new language. Most viewers watch muted; captions matter more than the dub.
6. Label the result as AI-modified where the platform requires it, and keep the original for comparison.

If you need shorter pieces: the 10-second limit

Lip-sync endpoints for single clips, such as Kling LipSync, take video of only 2 to 10 seconds. A 30-second Reel therefore needs three pieces, synced separately and rejoined. The cut points should fall at natural pauses in the speech, not mid-word.

Where translations slip

The model is rarely the problem; the text is. Idioms translate literally, numbers and dates change format, jokes lose timing, and names of products are mangled unless protected. Spanish and Russian run longer than English, and German longer again. Our notes on Spanish and Russian cover the specific traps, and what makes lip-sync dubbing look right covers the mouth.

What not to do

Do not translate and re-voice a video you do not have the rights to, and do not clone another person's voice. Putting translated words in a real person's mouth, in a voice that sounds like theirs, without their agreement is the use that gets accounts reported. The tool is neutral; the use is not. Our tools ask you to confirm the right to the source before they run.

A note on ChatGPT and similar assistants

An assistant can translate your script well, and cannot re-voice or re-time the video. That job sits in a dubbing model; what ChatGPT hands off lists the division of labour.

A worked example

Take a 30-second Reel in English where you talk to camera. Run HeyGen Translate into Spanish: $1.50. Have a Spanish speaker watch it with sound and tell you where it sounds unnatural; fix those phrases in the script and rerun only if the cost of a rerun is lower than the cost of posting something awkward, which for $1.50 it almost always is. Add Spanish captions, post it with a note that it is AI-dubbed, and compare its reach against the English original over a week. If it works, repeat for the next two languages in your audience.

Common questions

Do I have to re-time the lips?

If the speaker is on camera and the viewer will watch the face, yes. For voice-over on b-roll or a screen recording, a voice-only dub is enough and much cheaper.

What does a 30-second dub cost?

$1.50 with HeyGen Translate in one call, or $0.30 for the voice alone with ElevenLabs.

Why does my Spanish version look rushed?

Translations are often longer than the source. Tighten the script, or the mouth will race to fit the same seconds.

Can I dub someone else's video?

Only with the right to translate and re-voice it. A neutral tool does not make an unauthorised dub acceptable.

Should I add captions?

Yes, in the new language. Most viewers watch without sound.

How to do this without code: The whole translate, re-voice and re-time chain is one upload. In AutoHustle this is the Revoice tool: upload the video and pick the target language (about 1 credit per second of video, quoted before you press Generate). See How to revoice a video into another language and AI video revoice.

Related reading: HeyGen Video Translation vs Cheaper Alternatives (August 2026) · HeyGen vs sync for Lip Sync: $6, $5 and $8 a Minute Compared

For developers: the fal endpoints

Model	fal endpoint
HeyGen Translate v2 (precision)	fal-ai/heygen/v2/translate/precision
ElevenLabs Dubbing	fal-ai/elevenlabs/dubbing
Kling LipSync audio-to-video	fal-ai/kling-video/lipsync/audio-to-video
sync-3 lipsync	fal-ai/sync-lipsync/v3

HeyGen $0.05 per second; ElevenLabs $0.60 per minute; sync-3 $8 per minute; Kling LipSync $0.014 per undefined unit. Read on 18 October 2026.