Livestream service intro: 10 seconds of Wan 3.0, 63 cents to $2.50

A 10-second livestream intro on wan-3.0 costs $0.63 at 480p, $1.25 at 720p and $2.50 at 1080p on Sume. An intro that plays full screen on a 1080p stream every week is the case where the $2.50 tier is worth it, because you pay once and reuse the file.

Reuse is the point. A bumper is generated once, saved, and placed at the top of every service or stream.

Why 1080p fits this job

A weekly stream is almost always encoded at 1080p, and an intro at 720p is upscaled and softer next to your 1080p camera feed. Since you generate it once, the extra $1.25 over 720p is spread across every broadcast. Wan 3.0 supports 480p, 720p and 1080p, from 2 to 30 seconds (video generation docs).

A workflow that costs under $5

Run the prompt at 480p first, using your own logo or building photo as a first frame. When one take has the motion you like, re-run the same prompt and image at 1080p. Two 480p drafts plus one 1080p final is $3.76.

• Draft 1 and 2 at 480p: $0.63 each.
• Final at 1080p: $2.50.
• Add the title text in your streaming software, not in the model. Generated text is unreliable.

Audio and music

The model can generate audio, but a stream intro usually uses licensed or in-house music. Set generate_audio to false and lay your own track over it. If you want the clip to follow a track, wan-3.0 accepts an audio reference as an input_references entry of type audio_url.

Caution

Do not generate recognizable real people from a prompt. Use photos you have the right to use, and review the result for anything that could misrepresent your community. The billed amount is in usage.cost on the poll response (jobs and results).

Practical notes

Seasonal variants are cheap once the look is set. Keep the same first-frame image and change only the light and color words in the prompt: warm candlelight for winter, morning light for spring. Each variant at 1080p is $2.50.

Store the final file with the prompt and the settings used. Sume jobs return a file, and the prompt is what lets you recreate the same style next season without guessing.