TikTok ad language rule: when a video needs subtitles
TikTok's ad policy requires subtitles in an accepted language when the ad's language is not accepted for the market. How to burn them with Sume Video captions.

Yes. TikTok's ad format policy says the language, currency and other elements of an ad and its landing page must be consistent with the ad setting or acceptable for the target market. If the setting language is not specified, the ad content and landing page must be in an acceptable language for that market. Otherwise, subtitles in a language acceptable to the target market must be included. So an English-voiced AI ad going to a market that lists only another language needs burned-in subtitles. Sume's Video captions burns authored text at times you give.
What does the policy list per market?
The page carries a list of acceptable languages by market and a note that market-specific exceptions and restrictions may apply, with a pointer to a TikTok sales representative. A footnote says that for Japan the acceptance of language, currency and other elements is enforced. The table shows a few entries as the page printed them; use the page for your market, not this sample.
| Market | Acceptable languages listed |
|---|---|
| Australia | English |
| Bangladesh | Bengali, English |
| Cambodia | Khmer, English |
| Indonesia | Bahasa Indonesia, English |
| Israel | Arabic, Hebrew, English, Russian |
| Japan | Japanese (enforced, per the page footnote) |
What else does the language rule touch?
The same section says spoken language in the ad should ideally be clear, comprehensible and natural for the target market, that industries requiring disclaimers must present them in the local language, and that advertisers can consider a currency or region switch for currency consistency. A burned-in subtitle fixes the first case only if its wording is right, so check the translation, not only the timing.
How do you burn subtitles with Sume?
Video captions takes a public HTTPS video_url. With no wording supplied it runs speech-to-text and burns what it hears; language is only a hint to speech-to-text (for example ko or en) and never selects the style or font. To burn a translation you wrote or reviewed, pass cues (phrase-level) or segments with text, start and end in seconds. Authored cues skip speech-to-text and burn exactly that copy at those times. script_text, words, cues and segments are mutually exclusive.
The example below burns Indonesian, which is Latin script, with the slam style. The doc says an omitted style follows the wording, with slam for Latin and black-outline for Korean, and it names Hangul-specific styles such as korean-ad; I have not seen it document other scripts, so test a sample before you promise a market. A standalone job is $0.20 for videos up to 60 seconds, per the doc's pricing note.
curl -X POST https://api.sume.com/v1/video-captions \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: subs-id-001" \
-d '{
"video_url": "https://media.sume.com/artifacts/example/ad.mp4",
"style": "slam",
"cues": [
{"text": "Halo, selamat datang", "start": 0.5, "end": 2.0},
{"text": "Coba aplikasi baru kami hari ini", "start": 2.0, "end": 5.0}
]
}'Is a subtitle always the answer?
No. The policy's first route is to make the ad content itself acceptable: if the ad setting names a language, the ad and landing page should match it. Subtitles are the fallback when the ad is not in an accepted language. If you can re-voice the ad in the target language, spoken audio that is clear and natural for the market is what the page says it ideally wants, and burned subtitles then become an accessibility aid.
This post does not cover translation or dubbing, so treat translation as your own step: write or review the target-language lines, then burn them. For the exact fit between your market and an accepted language, the page tells you to ask a TikTok sales representative where exceptions apply.
How do you keep several markets under control?
Keep one master video and one cue list per market. Each market is a separate caption job on the master, so the voice track and picture are identical and only the burned text changes. If you only need a different look, pass source_caption_id with a new style: the doc says Sume reuses the source video and word timings, so no second speech-to-text runs.
Then review the result the way a reviewer will. Read each subtitle against the market's list in the policy table, confirm disclaimers are in the local language where your industry needs them, and make sure the burned text is legible on a phone.
Sources
Related posts
More in Use cases
- Will TikTok flag a brand's AI videos as AI spam? The July 2026 test
TikTok says it is testing detection for accounts posting AI spam on politics, financial advice and medical topics. What that covers and how to audit a series.
- TikTok Branded Mission Ad Specs: 9:16, 720x1280, 12-15 Seconds
TikTok's Branded Mission specs ask for 9:16 video of at least 720x1280, up to 100 MB, 12-15 seconds, over 2,500 kbps, plus MP3 music. Prepare it with Sume.
- TikTok catalog add-ons: 130px card, 546px tile, keep the bottom clear
TikTok's interactive add-ons take a 130px or 546px band at the bottom of the video. Check a clip with Sume video frames; lift captions with anchor_ratio.
- TikTok comment poll, two endings: branch one last frame for $11.56
Run a comment poll on how your episode ends. Branch two 10 second endings from one last frame with Seedance 2.5 for $11.56 in total, then post the winner.
Written by Sume