Asset Studio's four-step Omni workflow, rebuilt as Sume API calls
Asset Studio's flow is anchor, concepts, refine, export. Here is each step as a Sume call: /v1/videos, video-trim, timeline fit modes and a cropped square.

Asset Studio's four steps map onto four API moves: keep your brand brief as a fixed prompt block, generate concepts with POST /v1/videos, keep the good seconds with video-trim, and export formats through timeline fit modes plus a crop. Nothing here replaces Google's free tool inside Google Ads; it shows how to run the same shape from your own code when the clips must also leave Google.
Google's announcement (read 2026-10-04) names the steps: anchor brand guidelines, generate concepts, refine scenes, export multi-format. It also says assets get imperceptible SynthID watermarking and that the feature is free and rolling out globally across Demand Gen, Performance Max and YouTube. A companion page on multimodal video creation (read 2026-10-04) describes a path from brief to storyboard to final using Gemini, Veo and Nano Banana.
How does each step translate?
Sume has no brand-guideline object that I verified in the docs, so step one stays in your code. The rest map to documented endpoints.
| Asset Studio step | Sume call | Documented limit |
|---|---|---|
| Anchor brand guidelines | A fixed prompt block you prepend to every request | You hold it; no stored brand object |
| Generate concepts | POST /v1/videos, model gemini-omni-flash-1.1 | 3 to 10 seconds, 360p to 4K, 16:9 or 9:16 |
| Refine scenes | Re-run with an edited prompt, then POST /v1/video-trim | Trim is $0.02 per job, 0.2 to 900 s cut |
| Export multi-format | Timeline output 1080x1920 or 1920x1080, fit cover, contain, stretch or blur | $0.10 per output minute, rounded up |
How do I generate concepts without overspending?
Billing on /v1/videos is the provider list price times 1.25, reserved when you submit, and usage.cost on the completed job is the billable amount. GET /v1/videos/models lists per-SKU pricing, so read it before you queue a batch. Omni offers 360p, 720p, 1080p and 4K, so draft at a lower resolution and re-render only the winners.
Omni always returns audio and rejects generate_audio set to false, so plan for a soundtrack in every concept.
How do I keep only the good seconds?
A 10-second concept rarely works start to end. Send video-trim a start plus an end or duration, keep precision at exact, and leave audio on keep. The source must be a media.sume.com artifact, which a Sume render already is, and each request needs an Idempotency-Key.
How do I export every format from one master?
Timeline takes 1 to 200 video slots over an audio spine and accepts output sizes from 256 to 2160 on even integers, with a default of 1080x1920 MP4. The fit mode decides what happens when a clip does not match the frame: cover crops, contain letterboxes, stretch distorts, blur fills the bars with a blurred copy. Transitions of up to one second are available between slots.
For the square, crop as shown in the Omni square post. For storyboard planning before you spend on video, see the animatic recipe.
Where does this beat the free tool?
When one run must feed several channels, or when you want variants produced by a script and logged with job ids. The related post on bulk hook variants goes deeper. Inside Google Ads alone, the built-in tool is simpler and costs nothing per asset.
Sources
Related posts
More in Developers
- Google's June 15 deprecation notice gave 15 and 63 days: run a drill
Google announced Veo and Imagen 4 deprecations on Jun 15, 2026 with shutdowns Jun 30 and Aug 17. Here is a five-step drill that fits inside the shorter window.
- GPT-6.1 Sol function tool for Sume runs: clamp the cap in code
A Responses API function tool that lets GPT-6.1 Sol start a Sume Agent Completion: call_id as Idempotency-Key and a spend cap the model cannot exceed.
- GPT Image 2.5 widest banner: 3072x1024 at 3:1, with output cost
GPT Image 2.5 stops at 3:1. 3072x1024 passes the multiples-of-16 and pixel rules; 4:1 and 8:1 banners need another model. Output estimates by quality.
- GPT Image 2.5 returns base64; Sume returns a URL: port the code
OpenAI's image API returns base64 data for GPT Image models, and Sume's returns hosted URLs. The three lines that change when you move a decoder over.
Written by Sume