Vimeo AI dubbing in 29 languages vs Sume: one file per language

What does Vimeo's single-link dubbing mean for a Sume user?

Vimeo's AI page lists AI dubbing in 29 languages (availability varies by region), subtitle translation in 36+ languages, and single-link multi-language videos with no duplicate uploads. Sume has no equivalent container: each language you produce is its own finished file with its own URL on media.sume.com. The language picker is something your player, page or platform has to provide.

That is a real difference in where the work sits. With Vimeo the viewer's player swaps audio or subtitle tracks under one video. With Sume you generate N outputs and decide how to present them, for example one page per locale, one embed per language, or an upload to a platform that supports multiple audio tracks.

What else does Vimeo's page promise?

It lists voice replication so the dub sounds like the speaker, editable subtitles, and a note that translation typically takes about the length of the video. It also says you can review and edit all AI-generated content before publishing. AI features are included with Standard, Advanced and Enterprise plans, with extra AI credits available to buy; specific prices are not on that page, so none are quoted here.

Sume's counterparts are separate and narrower. You choose a TTS voice rather than cloning the speaker on the dub. Review happens in the gap between jobs, because each job returns a durable file or text you inspect before you spend on the next step. Sume jobs are asynchronous, and you can have a webhook call you when each finishes (see jobs and results).

How do you ship several languages from Sume?

Keep one clean master with no text burned in, then fan out. For captions, send one video captions job per language with authored cues (text, start and end in seconds), which skips speech-to-text and burns exactly your translated copy. Give every request a stable Idempotency-Key that includes the language so a retry never double-bills.

The script below submits Spanish and German jobs together and prints the job ids to poll. Replace the cues with your approved translations, and see three languages for $0.60 for the pricing shape.

import asyncio, os
import httpx

VIDEO = "https://media.sume.com/artifacts/artf_demo/clean.mp4"
CUES = {
    "es": [{"text": "Bienvenidos al curso", "start": 0.0, "end": 2.5}],
    "de": [{"text": "Willkommen zum Kurs", "start": 0.0, "end": 2.5}],
}

async def caption(client, lang, cues):
    r = await client.post(
        "https://api.sume.com/v1/video-captions",
        headers={"Idempotency-Key": f"course-intro-{lang}-v1"},
        json={"video_url": VIDEO, "cues": cues},
    )
    r.raise_for_status()
    return lang, r.json()["request_id"]

async def main():
    headers = {"Authorization": f"Bearer {os.environ['SUME_API_KEY']}"}
    async with httpx.AsyncClient(headers=headers, timeout=30) as client:
        jobs = await asyncio.gather(*(caption(client, k, v) for k, v in CUES.items()))
    print(dict(jobs))

asyncio.run(main())

Subtitle track or burned-in captions?

Vimeo's translated subtitles are a track viewers can switch on and off. Sume's caption job burns the text into the pixels, which is why you need one output per language. The upside is that the captions look the same on every platform and survive reposting; the downside is there is no off switch and no accessibility track. If you need a selectable track, our post on hard vs soft subtitles covers when to keep a sidecar file in addition.

Sume's caption language field is only a speech-to-text hint. It never picks the style or the font, so a Latin style renders Spanish and German fine while Korean copy needs a Hangul style. The docs describe Latin display faces and a set of Hangul faces; they do not list other scripts, so test any non-Latin, non-Hangul language before you commit to a batch.

What about dubbed audio?

For voice, run TTS per language with language set, then either replace the audio inside a Timeline 1.0 render (audio.url or audio.parts[]) or mint a reusable file with timeline audio. Each result is another standalone MP4 or audio file, not a track attached to the original.

When is Sume the right fit?

Choose Vimeo when your audience watches inside Vimeo and a single embedded player with a language menu is the goal. Choose Sume when the output is an ad, a short or a landing-page clip, where you want a finished file per market and an API that fits into a build script.