Board-game crowdfunding trailer from box, minis and cards, Wan 3.0

Send up to 10 reference images of box, miniatures and cards to wan-3.0 for a 15-second trailer: $1.88 at 720p on Sume, with logo and card-text drift.

5 min readSume
All posts

Send ten photos of the game, such as the box front, three miniatures and six cards, as input_references to wan-3.0 and ask for a 15-second trailer. On Sume that is 15 x $0.10 x 1.25 = $1.875, rounded up to $1.88 at 720p. Expect the box shape and miniature silhouettes to carry across the clip better than any printed text, because the vendor lists on-screen text accuracy as still improving.

Pick the ten references

  • One straight-on box front and one angled box shot.
  • Three miniatures, photographed against a plain background at the angle you want to see them.
  • Four card faces chosen for art, not for rules text.
  • One shot of the board or play area.
Wan 3.0 trailer length against price on Sume (read 2026-10-07)
Length480p720p1080p
15 s$0.94$1.88$3.75
20 s$1.25$2.50$5.00
30 s$1.88$3.75$7.50

Prompting without tag syntax

Morphic's page says labelling each reference in the prompt matters as much as supplying it. Sume's docs describe references as visual guidance and do not define a tag or numbering syntax, so describe each reference by what it looks like: the dragon miniature with the red base, the box with the silver crest. A trailer prompt can run as a short arc: box on a table, camera moves to the miniatures, cards fan out, hands move pieces on the board.

Wan 3.0 accepts 10 reference images on Sume, so a game with more components needs a selection or a second clip.

Where the text will be wrong

Card text, rule numbers and the logo on the box are the parts most likely to be redrawn. Plan for it: keep the generated clip text-free by asking for blank card backs and no captions, then place the real title card, rewards and campaign URL in an editor. Sume's timeline docs take Sume-hosted clips, so the end card can be joined to the clip afterwards.

Check each miniature against the photo. A backer will notice if a figure's weapon changes hands.

Trailer structure for a 15-second cut

Fifteen seconds is enough for three beats of five seconds. Beat one is the box on a table with the camera pushing in. Beat two is the miniatures lined up, then one stepping forward. Beat three is the cards and board, ending on a wide shot with room for a title overlay. If the campaign wants a 30-second spot, the price doubles to $3.75 at 720p, and the extra time needs more story, not more products.

Wan 3.0 generates audio by default in the catalog entry (generate_audio is optional), so decide before submitting whether you want the model's sound bed or your own music. A trailer usually wants its own track and the real voice-over, so set generate_audio to false when you plan to replace it and avoid choosing between two soundtracks later.

Run order

  • Draft at 480p (15 s is $0.94) and review the references that were ignored.
  • Fix the prompt, then render 720p ($1.88) for review.
  • Render 1080p ($3.75) only for the clip that goes on the campaign page.

Sources

Related posts

More in Use cases

All Use cases posts

Written by Sume