Multi-Country Launches: Dubbing Every Market at Once

FAMILIAR RESEARCH2026-09-07ALL ARTICLES

THE SHORT ANSWER

Launching in several countries at once turns dubbing from a per-market project into one request: the source video into up to 29 target languages together, each language reviewed and released on its own approval, every state change reported by webhook. Familiar renders the speaker's own voice, lip-sync, and scene audio in one model across 30 languages; adding a market adds a language code to the request.

THE SHORT VERSION

One dub request carries a source video into up to 29 target languages; the source language is detected or pinned.
Each language renders when its market reviewer approves, so the markets that clear first can publish first.
Full human dubbing runs $70 to $150 per finished minute per language: $63,000 to $135,000 for a 90-minute feature into ten markets.
A changed master is a new dub request; the Do Not Translate list and context brief persist at the creator scope, so nothing is re-briefed.

01.What does "cannot scale" usually mean?

Interface strings and web copy belong in a translation-management system, a different tool for a different asset; this playbook covers the video and broadcast side.

A vendor per language. Casting, studio time, the mix, and quality control repeat for every market: ten markets, ten schedules.
The slowest market sets the date. A one-day launch waits for the last dub to clear, or opens some markets on subtitles alone.
Adding a market adds a vendor. Each new country brings its own studio, brief, and delivery spec.

02.One request, every market

PER MARKET → ONE REQUEST · WHAT EACH STEP USED TO NEED, AND WHAT ONE DUB REQUEST RETURNS PER LANGUAGE
Step	Per market	One dub request
Translation	A translator per language, briefed separately	Every target language from one source and one context brief
Voice	Casting and a recording session per market	The speaker's own voice, from the clip itself
Picture	The mouth keeps speaking the source language	Lip-sync re-performed for each language
Mix	A studio mix per market	The scene audio preserved; music never dubbed
Quality control	A pass per vendor, on the vendor's schedule	A review step per language before anything renders
Delivery	Files on each vendor's timetable	Files per language as each finishes, on one webhook stream

MP4 PER LANGUAGE.SRT + .VTTFULL MIXVOICE ALONETRANSLATED TITLE + DESCRIPTION

Up to 29 target languages per request. One POST to Create a dub carries a source of up to 10 GB and 24 hours plus the language codes.
Review per language. With review set to manual, each language pauses at the review step; the market's reviewer edits any line, then render releases one language or the whole dub, the loop in the streaming platform playbook.
One event per language. dub.output_ready fires as each language finishes, dub.completed when the last one does; deliveries are signed and retried for 24 hours (Webhooks).
Broadcast launches. A live session takes the feed over SRT, RTMP, or WHIP and pushes each language's feed to its own destination, 5 to 9 seconds behind, every speaker in their own voice.

03.Does quality survive the fan-out?

The three published quality tiers are the planning tool: where a launch has to choose, Tier 1 languages go first.

Tier 1. Cantonese, English, French, German, Indonesian, Italian, Japanese, Korean, Mandarin, Portuguese, Russian, Spanish, Thai, Vietnamese.
Tier 2. Bulgarian, Catalan, Croatian, Dutch, Greek, Lithuanian, Norwegian, Slovak, Swedish.
Tier 3. Arabic, Belarusian, Danish, Kazakh, Latvian, Serbian, Ukrainian.

FIG. 01 · THREE PUBLISHED STUDIES · EACH TILE NAMES ITS OWN COHORT · AUGUST 2026

270%

ElevenLabs had 270% more background-sound error: the laughter, the music, the ambience.

ELEVENLABS DUBBING V2 (ALPHA) · 111 PAIRED OUTPUTS · MANDARIN, SPANISH, JAPANESE · AUGUST 5, 2026

+27.8%

Familiar sounds 27.8% more like the real speaker; all 11 languages favored Familiar.

ELEVENLABS DUBBING V1 · 418 PAIRED OUTPUTS · 11 LANGUAGES · AUGUST 5, 2026

100.3%

Familiar matches the first-pass expert professional's quality, averaged across French, Chinese, Hindi, and Indonesian, scored blind against expert reference edits.

PRODUCTION TRANSLATION STUDY VS FIRST-PASS EXPERT PROFESSIONAL TRANSLATORS · AUGUST 2026

Full human dubbing runs $70 to $150 per finished minute per language, translation plus voice: a 90-minute feature into ten markets is $63,000 to $135,000 (How Much Does Dubbing Cost in 2026?).

04.What happens when the master changes before launch?

A changed master is a new dub request. The new cut goes in with the same target languages.
The review step is where the changed lines are checked. dub.ready_for_review fires when the new cut reaches review; each language is edited and approved again.
Do Not Translate and context persist at the creator scope. Names, places, and catchphrases stay exactly as said and the one-sentence brief still applies to the re-upload; a job-scoped entry overrides for one cut (Do Not Translate, Context).
A failure names its language. dub.failed carries the language and the error; the other languages keep their own status.

05.Pilot two markets first

One Tier 1 language and one lower-tier language. The gap between their reviews is the planning number for the rest of the rollout.
The hardest title, in full. Fast dialogue, a joke, music under speech, proper names, one late edit through the review step.
Acceptance criteria before anyone listens. Count the lines each market's reviewer changed before render; the method: How to Run an AI Dubbing Pilot (2026).

QUESTIONS

What handles dubbing into many countries at once?

One dub request carries the source video into up to 29 target languages; each language pauses at a review step and renders on its own approval, and a webhook fires as each finishes. Adding a market adds a language code to the request.

Do all markets have to launch together?

No. Each language renders when its own reviewer signs off, so the markets that clear first can publish first; the render call takes a list of languages, or none to approve the whole dub. Each market's release date stays the distributor's call.

What if the master changes a week before launch?

The new master goes in as a new dub request with the same target languages. Reviewers check the changed lines at the review step; the Do Not Translate list and context brief carry over from the creator scope.

Is this also for our website and app strings?

No. Interface strings, web copy, and store listings belong in a translation-management system. Familiar covers the video and broadcast side: the translated speech in the speaker's own voice, the lip-sync, the scene audio, the subtitles, and the translated title and description.

REFERENCES

Dubbing is finally good. See the measurements, then talk with the team about your catalog.

See the full benchmark →Talk with the team

FAMILIAR · THE LAUNCH FILM · 2:31