Google Vids makes free Omni 1080p scenes: when to use an API

Google Vids now generates video scenes with Gemini Omni 1.1 Flash at no cost for anyone with a Google or Google Workspace account, per Google's September 23 post. If you need one clip for a deck, use Vids. If you need many clips from a script, a fixed size, or a file that lands in your own storage, call the model through an API such as Sume's.

What did Google add to Vids?

Google's post says anyone with a Google or Google Workspace account can generate high-quality videos at no cost with Gemini Omni 1.1 Flash, from vids.new on desktop. It lists 1080p HD scenes, extending scenes with smooth transitions, setting specific clip durations, upscaling existing AI clips, ready-made templates and a SynthID digital watermark in the video frames. It also says Workspace Business and Enterprise plans get expanded generation pools, and that text-to-speech narration in more than 100 languages, via Gemini 3.8 Flash-Lite, is coming soon.

The post gives no daily clip count and no regional list, so I cannot tell you how many scenes a free account makes. Treat "no cost" as the plan price, not a promise of unlimited volume.

Where does an API still win?

Vids is an editor. It does not give a script a request id, a retry key, or a webhook. The Gemini API page lists Omni Flash as paid-tier only with no free tier, at $17.50 per million video output tokens, which Google's pricing page converts at 5,792 tokens per second of 720p video, about $0.10 a second.

Sume lists the same model as gemini-omni-flash-1.1 and bills provider list times 1.25 per output second by resolution, so 720p is roughly $0.125 a second before any rounding. What you buy for that margin is the job envelope: an idempotency key, a status URL, a durable result URL and a catalog you can switch models in.

What does the API call look like for the same scene?

On Sume, an Omni text-to-video job is one request. The Video Router docs list 3 to 10 seconds, 360p to 4K, and 16:9 or 9:16, with native audio always on.

Keep the prompt concrete and tell the model the camera and lighting, as you would in a Vids prompt box. The result comes back as a job you poll, as described in Video generation.

curl -X POST https://api.sume.com/v1/video-router/generate \
  -H "Authorization: Bearer $SUME_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: scene-intro-001" \
  -d '{
    "model": "gemini-omni-flash-1.1",
    "prompt": "A slow push-in on a whiteboard as a team lead points to a roadmap, soft office light",
    "resolution": "1080p",
    "duration": 8,
    "aspect_ratio": "16:9",
    "mode": "async"
  }'

What does Sume not give you here?

Sume has no editor timeline UI inside this call, no Vids templates, and no narration step bundled in. Vids' Workspace admin controls are also outside what Sume offers. Sume's docs say nothing about extending a Vids clip.

Pick by volume: one to five scenes for a presentation, Vids. A repeatable pipeline, Sume or the Gemini API directly. The comparison of Google surfaces in which Google surface for client work covers the account side.