HeyGen translates 10 languages at once: what Sume does

The comparison in one paragraph

HeyGen's translation takes one video and returns it in up to 10 languages per job, with a dubbed voice and optional subtitles. Sume does not translate or dub. It burns text onto a video you already have, one job per clip, and the Video captions docs price each accepted standalone job at $0.20 for videos up to 60 seconds under the current fixed estimate, to be confirmed in GET /v1/catalog. If you already have the translated text, ten languages means ten jobs and $2.00 for a clip within that length. If you need the translating and dubbing done for you, that is a different product.

What HeyGen's pages say

The help article (read 2026-10-04) says one job can target up to 10 languages and lists credits per minute: Audio Only 4, Speed 6 and Precision 10. The developer skill page (read 2026-10-04) says billing is source duration multiplied by language count, and that a typical job takes 5 to 15 minutes, with some over 30. Its translator page (read 2026-10-04) lists 175+ languages, proofreading and AI subtitles.

Taking the help page figures at face value, a one-minute clip on Precision into 10 languages comes to 100 credits, by the multiplication the developer page describes. The credit-to-dollar rate depends on your plan, which those pages do not give, so this post stays in credits.

What Sume does per language

Sume's caption language field is a speech-to-text hint only. It never selects the style or the font, and it does not translate. For each target language you supply the words yourself: script_text, words, cues or segments, which are mutually exclusive in a request. Cues and words skip speech-to-text and burn exactly the copy you provide at the times you give.

Mind the script limits: the Latin styles slam, punch and tiktok-green reject Korean text with 400 caption_hangul_text_latin_style, so a Korean version needs one of the Hangul styles. SRT uploads are unsupported; convert the SRT into cues first.

Side by side

The numbers below are the documented ones from each side.

When to pick which

Choose the translation tool if you need the audio itself in new languages. Choose Sume captions if your audio is already right, such as an avatar clip that speaks one language, and you want styled text in others over the same picture. The three-language caption post shows the loop, and the SRT versus burned-in post covers the output formats. Reuse the same Idempotency-Key pattern per language-and-clip pair so a retry does not render twice.