GateTrue

What each generator documents about marking its output

Wan hands the decision to whoever sends the request

Wan's API reference documents a boolean watermark field defaulting to false, so a caller decides. The guide that once described a console stamp no longer says anything on the point. From 2 source pages, read 2026-10-25; no maker pays to be listed.

Who holds the marking decision on this rowThrough the documented interface the marking decision is a boolean in the request, so it belongs to whoever writes the call. The pages read publish no other control.Who holds the marking decision on this rowA caller writing a requestAn integratorA boolean field with a published defaultof false. Changed by editing one line.A person using a settings screenNot documentedThe pages read publish no switch for aperson, and no plan attached to the mark.Enforced by whoever last edited the call, if by anybody
Fig. 1 A clean file from this model carries no indication of whether anybody made a decision at all.
One cell of the register, and what it rests on. Read from the vendor's text-to-video API reference on 2026-10-25.
The readingWhat it holds
The recorded valueA request parameter, not a plan
The vendor's own wordsA boolean watermark parameter, default false, and no other control published.
How the reading was takenRead from an API reference for the parameter; the video generation guide now says nothing about a switch.

Inclusion rule. One row for the value, one for the sentence or absence behind it, one for the method. Nothing here is drawn from output, support replies or coverage by anyone else. Order. Fixed order: the value, then its evidence, then the method.

1Control belongs to a role, not to an account

On the plan-gated rows the person who can change the mark is the person who pays. Here it is the person who writes the request, which in most organisations is not the same person and often not the same department. An integrator can decide what every file carries without touching a subscription.

That is a different kind of exposure. A marking policy enforced through a plan is enforced by procurement; a marking policy enforced through a request field is enforced by whoever last edited the call. Neither is stronger in general, and the second is easier to change and harder to notice.

2What the register no longer records

The reading made on 2026-09-22 also held a statement from the video generation guide about output made on the vendor's own generation page. That page now redirects to a general overview of video generation and editing, which carries no such statement, so the register records the console side of the control column as not documented instead of keeping a line the vendor no longer prints.

The API reference is unchanged on the point: the field is there, and its default is false.

3What a default of false implies for anybody downstream

Nothing is applied unless a caller asks for it, which reverses the usual question. Elsewhere this column asks what it takes to remove a mark; here it asks who would add one and why. The reference gives no reason either way, and leaves the choice to the caller.

For a receiving party the practical consequence is that a file from this model carries no information about whether anybody made a decision. A clean file might have come from a caller who set the field to false or from a caller who never knew the field existed, and those are different situations with the same artefact.

4From the production side

Next to the marking question in “Wan, control”, SceneMixer's pages describe its output like this:

Recommended is the badge on the tier the site suggests for most character sheets; the default for new accounts stays the cheaper tier.

SceneMixer, Qwen-Image 3.0 Pro: the recommended tier — “Is Recommended the same as default?” (read 2026-10-13)

Dubbing replaces the audio of a finished clip; the picture was generated for a different line, so mouth shapes, pauses and shot length all belong to the original language. Native-language generation gives the video model the line in the target language before it renders the shot, so the performance, the pacing and the cut points are built for that language.

SceneMixer, Native-Language Dialogue vs. Dubbing in AI Video (2026): Why the Cast Should Speak the Line (read 2026-10-13)

5Sources

  • Paid tier
    The API request has a boolean watermark parameter that defaults to false; set true it adds a watermark in the lower-right corner with the text AI Generatedthe API surface, default offModel Studio, text-to-video API reference / since 2026-09-22 / checked 2026-10-25
  • Control
    Not documented on the vendor's public pages (as of 2026-10-25)nothing published about a console switch: the earlier statement is no longer on the pageModel Studio, video generation and editing guide / recorded 2026-10-25

The value above is read from alibabacloud.com, the vendor's text-to-video API reference, on 2026-10-25. The whole row for this generator is on its own page, and the rule for what this column may hold is on the control column page. Neighbouring readings: Luma AI, control, Sora, control.