TikTok streaming ads specs: 2:3 card image and 35-character title
TikTok's streaming ads page lists a 2:3 card image at 500x750 or larger and a two-line, 35-character title; video follows in-feed rules. Make the card in Sume.

TikTok's streaming ads page asks for a card image at 2:3 and at least 500x750, a title of up to two lines and 35 characters, and video that follows the in-feed guidelines. Sume can produce the card with a 2:3 image request. Read 2026-09-29.
What does TikTok list?
Video is not given its own table; the page points to in-feed rules (9:16 at 540x960 or larger, per the auction page).
| Item | Value |
|---|---|
| Card image ratio | 2:3 |
| Card image min size | 500x750 |
| Title | 2 lines, 35 characters |
| Video | Follows in-feed guidelines |
How do I generate a 2:3 card?
2:3 is in the image docs' ratio list. With openai/gpt-image-2.5, use image_size 1024x1536: both edges multiples of 16, 1,572,864 pixels (inside 655,360 to 8,294,400), and exactly 2:3, above the 500x750 minimum.
curl -X POST https://api.sume.com/v1/images \
-H "Authorization: Bearer $SUME_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "openai/gpt-image-2.5",
"prompt": "Streaming series key art, lone astronaut on a red dune, clear space at bottom for a title",
"image_size": "1024x1536"
}'Should text go in the image?
The page treats the title as a separate field, so leave the bottom clear rather than having the model draw text. Image models can misspell text; add the title in the ad tool.
What are the limits of this answer?
Sume does not submit streaming ads. Check TikTok's page for the current specs before you deliver.
Sources
Related posts
More in Use cases
- TTS API for video narration: script to voice track to MP4
Narrate a video with Sume's TTS API: send up to 20,000 characters, get a voice file, then lay it under clips with Timeline 1.0.
- Remix a viral TikTok video with AI: what Sume does and does not do
Trending search returns TikTok metadata and watch URLs, not video files. To remix, ingest a clip you own, read its shots and text, then generate new footage.
- Virtual twilight: turn a daytime exterior into a dusk photo
Virtual twilight turns a daytime house photo into a dusk shot with a sunset sky and lit windows. How to make one with AI, and what to check after.
- Word document to video AI: turn a .docx into a narrated video
To turn a Word document into a video, extract its text and images first. Sume's agent takes text in input and images as attachments; a .docx isn't one.
Written by Sume