Pan de muerto bakery clip in English and Spanish, Nov 1-2

To promote pan de muerto for Day of the Dead, animate one photo of the loaves with Wan 3.0 for 6 seconds at 480p ($0.375 on Sume), then run two caption jobs on that clean clip, one with English cues and one with Spanish cues, at $0.20 each. That gives you two ready-to-post versions for $0.775 in total, with no speech to transcribe and no voice-over to record.

The holiday is traditionally celebrated on November 1 and 2, according to the Day of the Dead article on Wikipedia (read 2026-10-11), which also describes pan de muerto as a sweet roll topped with sugar and often decorated with bone-shaped pieces. In 2026 that is a Sunday and a Monday, so the useful posting window for a bakery is the week before: orders for the weekend close first.

What do I need before I start?

One photo of your own bread, shot in daylight, plus a short written offer in two languages. Use your own photo as the first frame instead of asking a model to invent the loaf: customers want the bread they will actually pick up. Keep marigold petals, candles or a plain tablecloth in the frame if you like, because the model animates what is there.

Sume's video generation docs (read 2026-10-11) describe a frame_images field for image-to-video, with the image sent as a first_frame. They list Wan 3.0 as accepting 2 to 30 seconds. The prompt can stay modest: slow push-in, steam rising, a few petals drifting. A short, calm move reads as respectful and survives the compression of a feed.

How do I add English and Spanish text without speech?

The video captions docs (read 2026-10-11) say speech-to-captions only works when the clip has audible speech, and a silent clip fails with caption_no_speech. For a silent clip you send cues instead: each cue has text, start and end in seconds, and Sume burns the text at those times without running speech-to-text.

Run the job twice on the same clean clip, once per language. Do not stack both languages in one job unless you want both on screen at once; two separate files let you post the Spanish version to your Spanish-language audience and the English one to the rest. Pick a style explicitly. The docs say a Latin script defaults to the slam style, and you can choose another style by name. Before you post, check a frame for accented letters such as the o in muerto and the n in anos or an inverted question mark if you use one, and re-run with a different style if a glyph drops.

What does the whole job cost?

Wan 3.0 is billed per output second by resolution. At 480p the Sume rate is $0.0625 a second, so 6 seconds is $0.375. The caption job is a flat $0.20 for a video up to 60 seconds. The numbers below come from the Sume rate tables in the repository and the docs page above; the live price is always in GET /v1/catalog.

What should the cues say?

Keep each cue to a few words and leave each on screen for about two seconds. A six-second clip fits three cues per language.

• 0 to 2 s: the product, such as Pan de muerto, baked fresh / Pan de muerto, recien horneado.
• 2 to 4 s: the order rule, such as Order by Friday, pick up Nov 1 / Pide antes del viernes, recoge el 1 de nov.
• 4 to 6 s: your shop name and street.
• Leave out claims you cannot back up, such as sold-out-by dates you have not set.

What should I check before posting?

Watch the clip once with the sound off, since most feeds autoplay muted, and once at phone size to confirm the text is legible over the bread. If the loaf drifts out of shape, regenerate with the same first frame; a six-second Wan clip at 480p is cheap enough to try three takes ($1.125) before you pay for the captions. Caption only the take you keep.