Gemini API paid vs unpaid data use: Omni prompts and uploaded clips
Gemini API terms: unpaid content may improve Google products and reach human reviewers; paid content does not. What that means for Omni edit uploads.

Under Google's Gemini API additional terms, unpaid services use your content to improve Google products and human reviewers may read it, while paid services do not use prompts or responses to improve products and log them for a limited time for abuse monitoring. For Omni, that matters most for edit and extend, where you upload a source clip. Use a paid project for anything you cannot share.
The terms in plain words
The Gemini API additional terms say that for unpaid services, Google uses submitted content and generated responses to provide, improve and develop products, and that human reviewers may read, annotate and process it, with a warning not to submit sensitive or confidential data. For paid services, Google says it does not use prompts or responses to improve its products and processes data under its data processing terms, while logging prompts and responses for a limited period for abuse detection.
This is a summary of the page as read on 2026-10-04; check the live page before you rely on it, since terms change.
| Service type | Used to improve products | Human review | Retention noted |
|---|---|---|---|
| Unpaid | Yes | Possible | Not specified here |
| Paid | No | Not for improvement | Limited-time logs for abuse monitoring |
Why edit and extend raise the stakes
Omni's guide says uploaded videos for edit or extend must be 10 seconds or shorter, and video references are limited to three clips of three seconds. A short clip can still be an unreleased product shot or a face. On an unpaid project that clip is input content under the unpaid terms.
The guide also notes that editing or extending uploaded videos is unavailable in the EEA, Switzerland and the UK, so a European team may be blocked before the data question even arises.
What I can and cannot say about Sume
Sume's docs describe the pipeline, not its upstream providers' data terms, so I will not claim anything about those. What the Video Router docs do say is that an edit request takes a video_url, cannot be combined with image or reference inputs, and bills output seconds. The API reference lists the routes, but the Sume pages I read do not state upstream retention terms for uploaded clips, so get that answer in writing before you send sensitive footage.
Practical rule: classify the clip first, then pick the route. Public marketing footage can go anywhere; unreleased footage needs a written answer from whichever vendor you pick.
Sources
Related posts
More in Developers
- Gemini API's $10 per 10 minutes limit: how many Omni clips fit
Gemini API spend limits are $10, $50 or $200 per rolling 10 minutes by tier. At about $0.10 a second that is roughly 9 to 197 ten-second Omni clips per window.
- Gemini CLI 0.62 MCP titles: reading Sume's tool names
Gemini CLI v0.62.0 formats MCP tool call titles as structured signatures. Sume tool ids are underscore names such as generate_image; dotted aliases map to them.
- Gemini CLI v0.63 plan execution in CI: gate paid Sume calls first
Gemini CLI preview v0.63.0 adds autonomous plan execution in non-interactive mode. Before unattended runs, gate Sume paid tools with dry_run and max_spend_usd.
- Headless Gemini CLI and MCP auth: use a Sume API key
A headless Gemini CLI run cannot finish a browser consent. Connect to hosted Sume MCP with an API key header instead, and keep paid calls bounded.
Written by Sume