Synthesia Video Translation: Dubbing vs 1-Click Translation

Synthesia translates video in two ways, and your plan decides which one you get. AI dubbing replaces the spoken audio of any video file or YouTube link with a cloned voice, on every plan (on Enterprise, a paid add-on an admin turns on per workspace), in 70+ languages on Basic, Starter and Pro and 140+ on Enterprise. It uses 40 credits a minute, or 80 with lip-sync. 1-Click Translation rebuilds a video made in Synthesia, script and on-screen text included, into 140+ languages, but only on Enterprise.^[1] So the verdict is simple. Self-serve teams dub. Teams that need translated on-screen text, a multilingual player or multi-language SCORM need Enterprise.

Dubbing and translation solve different problems

Dubbing works on the audio track. You upload an existing video or paste a YouTube link, choose target languages, and Synthesia transcribes the speech, translates it and voices it with a clone of the original speaker; when several people talk, each voice is detected and cloned separately. Nothing on screen changes. Titles, graphics and burned-in captions stay in the source language, because dubbing translates spoken audio only. That makes dubbing the right tool for webinars, filmed interviews, product demos and anything you did not build in Synthesia.

1-Click Translation works on the project itself. It applies to videos made in Synthesia, keeps the avatar, scenes and design, and translates both the script and the on-screen text, with an option to translate the script alone. Each language becomes its own version, saved as a draft in a translation folder. Those versions feed the multilingual player, where viewers switch language inside one video, and a SCORM export that includes every translated version. Translation, the multilingual player and SCORM are all Enterprise-only.^[2]

A working rule follows from that grid. If the source is a recording, dub it. If the source is a Synthesia project on a self-serve plan, you can also write the script in the target language yourself, since avatar voices cover 140+ languages and accents on every plan. You lose the translation folder and the language switcher, but on-screen text gets rewritten by hand in the same pass. For two or three languages, that manual route is often enough.

How to dub a video, step by step

1. Pick the source. Upload a video file or paste a YouTube link; only one link per session, but Shorts and region-restricted videos work.
2. Choose the target language from the list your plan offers.
3. Decide on lip-sync. Lip-sync is listed on every plan and doubles the credits.
4. Set the timing. Adaptive, the default, adjusts video speed slightly to fit the translation; Original keeps the source length and changes only the voiceover speed.
5. Read the credit estimate. A modal shows it from the duration and the lip-sync choice before anything is charged.
6. Check the result. Listen to every language before publishing, especially product names.

Self-serve plans dub directly. The Review transcript step, which lets you correct the transcription before dubbing at no credit cost, needs Enterprise transcript editing. Proofreading or retranslating a finished dub uses no extra credits, so a bad first pass costs time rather than credits. Pronunciation is the weak spot. Dubbing cannot set custom pronunciations, and Synthesia warns that specialist terms may not sound as expected, so listen to product names in every language before publishing.

Dubbing needs no consent from the speaker, unlike avatar creation or voice cloning, but the uploader must hold the rights to the content. Synthesia creates a personal avatar or voice clone only with a live consent recording from that same person, who must be of legal age; it refuses celebrities, public figures, deceased people and anyone who has not consented, and you cannot create a photo avatar of someone else even with their permission. Dubs also pass automated content review, which rejects political, adult and violent material, hate and copyrighted material used without rights.

What dubbing costs on each plan

Every self-serve plan includes 10 minutes of video or dubbing, 3 survey responses and 500 credits a month; after that, dubbing draws from the shared credit pool at 40 credits a minute, or 80 with lip-sync.^[3] Synthesia's pricing card puts the extra dubbing capacity at up to 16 minutes a month on Starter and up to 78 on Pro, on top of the free allowance, or 197 and 947 minutes a year on yearly billing. The bill lists the base dub and the lip-sync add-on as separate lines.

Starter is $19 a month billed monthly, or $14 a month billed yearly ($168 a year). Pro is $89 a month billed monthly, or $59 a month billed yearly ($708 a year), at its entry credit level. Both are per account with one editor. Because credits are one pool, a heavy dubbing month leaves fewer credits for avatar video, surveys and Syren; you can buy more or set an automatic top-up. On Enterprise, dubbing is a paid add-on an admin turns on per workspace, and it uses contract credits even though Enterprise video generation does not.

Credits and the monthly allowance reset every billing period without rollover, as our Synthesia pricing guide explains with the yearly math. On Basic, dubbed videos carry a Synthesia watermark and cannot be downloaded as MP4. Synthesia's free dubbing tool on its website dubs only the first minute, also watermarked. Either route lets you hear the voice quality in your language before paying.

Translating Synthesia projects on Enterprise

Translation starts from the finished project. You pick languages or upload your own XLIFF file, decide whether on-screen text is translated or only the script, and choose whether every version generates automatically. Without auto-generation, each draft must be generated by hand before sharing. Publishing is manual too. You publish all translations from the publish menu, and a version that needs its own schedule must be duplicated to detach it from the root video. In the Share panel, an Include translations toggle adds a language selector to the player.

The translation folder is load-bearing. Moving a version out of the folder breaks language switching in the multilingual player, and deleting the folder cannot be undone. Interactive question text and answers do not translate automatically, so quizzes need manual edits in every version. Overflowing translated text is scaled down to fit its box, while text without a defined box, such as in animations, is shortened but not resized. If a brand font lacks a character the new language needs, that character appears in a different typeface.

Keeping terminology and voices consistent

Consistency tools sit almost entirely on Enterprise. The Glossary sets fixed translations, do-not-translate terms and pronunciations, applies to new translations and dubs, holds one list per workspace and imports from CSV, but is not available in workspaces that opted out of LLM-driven translation; for dubbing, its first 900 terms also improve transcription, though its pronunciation rules have no effect on dubs. Editing the original or translated transcript is Enterprise-only, as is reassigning, creating and renaming speakers in a dub. The pricing matrix marks an option to override a cloned speaker voice as coming soon, so do not plan around it yet.

XLIFF serves teams that already pay human translators. Since October 2026, Enterprise users can dub with their own XLIFF translations and mix them with machine-translated languages in one run, but only for imported videos. In 1-Click Translation, XLIFF must be version 1.2, one file per language, with no mixing in the same run. SRT and VTT are never accepted. Teams weighing a second tool should know HeyGen dubs into 175+ languages and dialects on paid plans but does not translate on-screen text either; our HeyGen video translation guide covers it.

Dubbing at volume: bulk uploads and the Localization API

The app's bulk dubbing, added in October 2026, takes up to 10 uploaded videos per batch with shared settings for languages, lip-sync and timing, and shows one credit estimate for the batch. YouTube links and the Review transcript step are not available in bulk. For larger volumes, Synthesia points to the Localization API, announced in October 2026. It accepts presigned S3 video URLs and up to 20 target languages per request, and it works on every plan with enough credit, drawing on the subscription's balance.^[4]

One default catches developers. The Localization API lip-syncs by default, which costs extra credit, while the old dubbing endpoint did not unless asked; turn lip-sync off to localize audio only. That older /v2 dubbing API is deprecated with no removal date announced, so new integrations should start on the Localization API. Teams without developers can dub from a spreadsheet through Synthesia's Excel add-in, which runs on an API key. Admins can switch dubbing on or off per organization and workspace.

The bottom line

For most self-serve buyers, Synthesia video translation means AI dubbing: any recording, every plan, 40 credits a minute or 80 with lip-sync, with audio translated and on-screen text left alone. That covers webinars, demos and filmed training well. Enterprise earns its price when the job is a library of Synthesia-made courses that needs translated on-screen text, a language switcher, a shared glossary and multi-language SCORM. Before choosing, dub one real video into your hardest language and listen for product names.