Find micro-drama premises: TikTok trending search for $0.10 a call
One trending-videos search returns ranked TikTok metadata for a keyword, from 1 to 50 results, for $0.10. How to read it for premises, not to copy clips.
POST /v1/trending-videos/search returns ranked public TikTok video metadata for one keyword, with 1 to 50 results, for $0.10 a call. Use it to see which premises, hooks and captions are moving before you write an episode, not to copy anyone's footage. The results are metadata and links, and the endpoint does not download or mirror videos.
Micro-dramas were reported as gaining traction on TikTok by Metricool, so a keyword such as a genre or trope is a fair first query.
What you send
platform supports only tiktok. On production the query is required, up to 200 characters, and can be a brand, product, creator or keyword. window can be yesterday, this-week, this-month, last-3-months, last-6-months or all-time; a generic production query defaults to this-month. region is an optional two-letter country code.
| Field | Notes |
|---|---|
| query | Required, up to 200 chars |
| window | yesterday through all-time |
| limit | 1 to 50, default 10 |
| region | Optional two-letter code |
| summary_mode | none (default), metadata or transcript |
What comes back
Each entry has a public watch url, an optional cover_url, description, created_at, region, author.handle, metrics and relevance scores. summary_mode transcript currently returns metadata and an unsupported warning, so do not plan around transcripts.
A premise-research routine
Keep the routine to a few calls so the cost stays visible.
- Run three queries with limit 10 across
this-week: $0.30 in total. - Read the descriptions for repeated setups, such as a secret, a swap or a reveal in the last second.
- Write your own three line premises that borrow the structure, not the story.
- Check any shortlisted video by watching it on TikTok, since the endpoint gives metadata only.
What Sume does not do
The search does not tell you why a video works, and metrics are a snapshot. It also does not clear rights to anything: do not reuse another creator's characters or footage. Build your own lead and scenes, then generate them with the video models.
The rebuilt browse mode with no query is on for the dev environment only; production keeps the required query.
Sources
Related posts
More in Developers
- Five Sume video submit errors and which ones to retry
insufficient_credits, idempotency_conflict, queue_full, rate_limited and provider_capacity_exceeded each need a different response. What to do, in a table.
- Flask webhook receiver for Sume: verify sume-v1, refuse empty secret
A Flask route that verifies the Sume signature on the raw body, takes either rotation entry, checks the replay window, and will not boot without a secret.
- FLUX 3 bounding box to a mask_url: Python region edit on Sume
FLUX 3 Image boxes use [top, left, bottom, right] on a 0-1000 grid. Convert one to an RGBA mask with Pillow and run the region edit on Sume's GPT Image 2.5.
- FLUX 3's element table as app state: run it on Sume with job metadata
Keep a FLUX 3 style element table in your own app, build each Sume edit from it, and tag every job with the table version through the metadata field.
Written by Sume