Can you use an AI voice on TikTok Shop?

It depends on where the voice plays. In a video you post, TikTok Shop's US policy allows AI-generated dubbing and narration. In a LIVE, it prohibits AI-generated voices, because a LIVE must use real-time speech or sign language.

The TikTok rules come from two US Seller Center pages read 2026-09-29: Promotional Content Requirements for High-Quality Videos and LIVEs (09/17/2026) and AI-Generated Content Restrictions and Requirements (09/02/2026).

Where is an AI voice allowed and where is it not?

The two pages split cleanly by format.

Do I have to disclose an AI voice in a posted video?

TikTok says creators must disclose content that is entirely AI-generated or significantly edited with AI tools. Disclosure can be on-screen text, a watermark, a sticker, a note in the description, or the AI-generated content toggle on the post settings page. The page does not say whether a synthetic voice alone counts as a significant edit, so treat a fully synthetic narration track as something to label when you are unsure.

The same page says labeling content as AI-generated when no AI was used is a violation, so label what is true.

How do I make a narration track with Sume?

Sume's TTS endpoint is POST /v1/tts-1.0/generate. The API reference example takes a transcript, a voice chosen by id, and an output_format. The job runs asynchronously, so poll it for the audio file.

curl -X POST https://api.sume.com/v1/tts-1.0/generate \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: shop-narration-001" \
  -d '{
    "transcript": "This case fits a 13-inch laptop. Here it is in use.",
    "voice": { "mode": "id", "id": "'"$VOICE_ID"'" },
    "output_format": { "container": "mp3", "sample_rate": 44100, "bit_rate": 128000 },
    "mode": "async"
  }'

How do I put the voice under the video?

Timeline 1.0 takes the voice file as the audio spine: audio.url plus audio.duration_seconds, with your clips as ordered video[] slots, and renders one MP4. Both the spine and each clip's source_url must be Sume-hosted files, such as the output of an earlier Sume job; there is no public upload route for arbitrary local footage. In current code a Timeline render takes sound only from the spine and the optional soundtrack, so the voice file is the only speech in the output. Stills are allowed as slots but render as static holds, and TikTok Shop says posted promo videos must not use still, looping, or scrolling images through the entire video without featuring your face and the physical product.

What should I check before posting a video with a synthetic voice?

Read the script against the product page first. TikTok's AI policy lists size, color, material, quantity, bundle contents, features and performance as things AI must not alter, and a spoken claim is still a claim. If the narration says a case fits a 13-inch laptop, the product page has to say the same.

Then watch the finished MP4 with sound on. If the voice and the footage disagree, fix the script or the clip before you upload, because a label applied by the platform cannot be reversed.

Can Sume run a LIVE with an AI voice?

No, and TikTok would not allow it. Sume's docs describe files you generate and assemble, not a LIVE broadcast. The narration workflow above is for posted videos only.