Gemini Omni can't use a YouTube link as its source video
Google lists YouTube videos as an unsupported media source for Gemini Omni Flash. On Sume, video_url is a URI field; use a direct link to the clip file.

Gemini Omni Flash cannot take a YouTube video as a media source: Google's limitations list says so in one line, "Using YouTube videos as a media source is not supported." Pasting a watch-page link to edit, extend or reference a clip is the wrong input. Upload or host the clip file itself.
Source: Google's Gemini API: Generate and edit videos with Gemini Omni Flash, read 2026-10-02. For Sume, Video Router documents video_url as the edit source and reference_video_urls for references; the OpenAPI schema types both as URIs, and every example in the docs points at a file such as clip.mp4. The docs do not say a YouTube page URL is accepted, so do not rely on one.
What counts as a source here?
In Google's guide a source is any video you ask the model to edit, extend or reference. The same page sets other limits on it: uploaded input videos for editing and extension must be 10 seconds or less, video references work best with likenesses, and any audio in a video reference is ignored. Referencing across multiple videos is not supported.
How do I get a usable file link?
Use a file you have the right to use, hosted at a public HTTPS address that serves the video directly. Sume's docs say results come back as Sume media URLs, which are plain links to the file.
Here is what an edit request looks like, taken from the Video Router page:
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: video-router-edit-001" \
-d '{
"model": "gemini-omni-flash-1.1",
"prompt": "Replace the bottle with an apple. Keep everything else the same.",
"video_url": "https://example.com/clip.mp4",
"resolution": "720p",
"mode": "async"
}'Which fields take a clip on Sume?
Two fields take clips, and they do different jobs.
| Field | Job | Limit for Omni |
|---|---|---|
video_url | The clip to edit | One clip; cannot be combined with image_url, end_image_url or reference lists |
reference_video_urls | Condition a fresh generation | At most 3, each at most 3 seconds; addressed as <VIDEO_REF_0> |
What if the clip only exists on YouTube?
Use the original file you uploaded to YouTube, then trim it to the length the model accepts. Sume's docs do not cover downloading from YouTube, so any download step happens outside Sume. Rights and platform terms are yours to check; this post does not give legal advice.
Sources
Related posts
More in Models
- Gemini TTS voices: 150+ or 2,000+? Two sources, one test plan
Google's changelog says 150+ Gemini TTS voices; a tracker says 2,000+. Do not pick by count. A voice test plan, and how voices are selected on Sume.
- Gemini TTS limit: 16,384 output tokens at 25 per second is 10:55
Gemini 3.8 TTS is reported to cap output at 16,384 tokens with 25 audio tokens per second, about 10 minutes 55 seconds. Sume's TTS cap is 1,200 seconds.
- gpt-5.6-terra and gpt-5.6-luna on Sume: which still run, as what
Sume retired GPT 5.6 from every picker. Terra has no successor and runs as itself, Luna moves to GPT-6 Luna only where that is admitted, Sol moves to GPT-6 Sol.
- GPT-6.1 Sol effort on Sume: picker offers Low, Medium, High, plus Fast
OpenAI lists five effort values for GPT-6.1 Sol. Sume's picker exposes three, keeps effort off the model id, and prices Fast as an opt-in. What that means.
Written by Sume