Generate Your First Video
Before you start
Before you start, confirm that you have joined an active team project. If no project exists, read Create Your First Team Project.
The page needs at least one available video model. Depending on the mode you intend to use, prepare a video description, first-frame image, or source video, and confirm that estimated credits do not exceed the team's balance.
1. Select a project
Choose a valid team project under Project. A project is required. An episode is optional; selecting one carries that project and episode context into later steps.
If the project is closed, ask an eligible person to reopen it. If it is archived, select or create another active project.
2. Select a model and mode
Choose an available model. The page shows the resolutions, durations, aspect ratios, sound settings, and modes available for that model, so use the current options.
For a first video, start with Creative mode, called Reference Generation by some models: enter a video description, then add reference media if needed. First/Last Frame mode requires a first frame; the last frame is optional. If that mode cannot be selected, choose a model that supports it.
3. Enter a video description
Describe the subject, action, scene, camera, and desired visual change. When asset mentions are available, type @ and select a person or reference asset that you already added.
For example:
A cyclist rides from the left of the frame into the distance on a city street after rain. A low camera slowly follows, warm lights reflect on the wet road, continuous movement, realistic cinematic style.
Add reference media first, then type @ to select its card and explain its role in the description. Follow the field label to see whether a description is required; frame mode and some modes with reference media allow it to be omitted.
Do not type text that only looks like an asset mention. Select from the page suggestions so the intended asset is referenced.
4. Review media and estimated credits
Availability and limits for reference images, videos, audio, and person assets depend on the current model and mode. If a media section is not displayed, the current combination does not support it. After uploading, wait for media checks or source-video metadata reading to finish.
Before submitting, confirm that:
- the project is still active;
- all media required by the current mode is ready;
- estimated credits and generation count match your intent;
- the team has enough credits.
For a first run, set the generation count to 1 so the task and credit use are easy to verify.
5. Submit and confirm the result
Select Generate. The page moves to Preview, and the task may pass through queued and processing states. You can also reopen it later from History.
What you should see
Preview contains a playable video, and History retains the same task with its actual status. When the task supports download, asset import, or continued editing, the corresponding actions appear on the page.
Troubleshooting
- A mode or asset input is missing: confirm that the current model supports it, then switch models or prepare the media requested by the page.
- Generate reports missing content: check the project, video description, and the required first frame in First/Last Frame mode.
- The task fails or mentions cannot be restored: review the full error and adjust media or parameters before reuse. Select usable media again and insert mentions from the
@list; changing a number alone does not repair the reference.
Next steps
Understand Video Modes and Assets