SEELE TV / Prompt archive

Wan 2.0 prompts written as reviewable shot directions.

Wan 2.0 prompts includes six text-led and six reference-led production starters. Copy a direction or send it to the current Wan template, then confirm the visible version and supported inputs before attaching media or generating.

Video generation6 text + 6 referenceVersion verification

Popular prompts / Wan 2.0

All video prompts

12 directions
Text generationProduct16:95 sec

Precision product reveal

Create a five-second 16:9 product film for an unbranded stainless-steel travel mug. Begin on a locked three-quarter frame, move one soft highlight across the metal, then let one hand rotate the mug 30 degrees and stop. Preserve the exact lid, handle, proportions, finish, table, and light direction. No logo, text, liquid splash, new controls, changing geometry, or performance claim.
Start Creating Source reserved
Text generationBrand9:165 sec

Vertical café preparation

Create a five-second 9:16 scene in an independent café before opening. One adult barista places an unbranded ceramic cup under the espresso machine, starts one pour, and ends on the cup in warm window light. Use a restrained shoulder-height camera and preserve identity, apron, cup, machine, counter geography, and lighting. No trademarks, menu text, extra people, impossible steam, or copied music.
Start Creating Source reserved
Text generationGames16:95 sec

Original game-world reveal

Create a five-second original fantasy game-cinematic concept. A courier enters a ruined observatory and stops when a dormant brass guardian opens one eye. Track behind the courier, then settle into a wide composition that preserves costume, satchel, guardian geometry, room geography, screen direction, and dusk lighting. No franchise resemblance, HUD, logo, duplicated character, attack, or changing creature design.
Start Creating Source reserved
Text generationMotion16:95 sec

Physical camera-move test

Create a five-second continuous shot of a cyclist passing beneath a concrete overpass after rain. Start low beside a puddle, pan smoothly with the cyclist, and finish as the rider exits frame while the reflection remains stable. Preserve rider identity, bicycle geometry, road direction, wet-surface behavior, background traffic, and overcast light. No cuts, reversed wheels, duplicated limbs, sudden weather change, text, or logo.
Start Creating Source reserved
Text generationEcommerce1:15 sec

Ecommerce material close-up

Create a five-second 1:1 ecommerce concept for an unbranded olive canvas pouch. Show one hand placing a folded map inside and closing the zipper once, ending on a stable material close-up. Preserve pouch dimensions, stitching, zipper, hand identity, map, stone surface, and daylight. Treat this as an illustrative material view, not proof of durability. No logo, extra pocket, changing color, impossible fabric, text, or certification claim.
Start Creating Source reserved
Text generationDialogue16:95 sec

Quiet dialogue reaction

Create a five-second original dialogue reaction in a late-night laundromat. Two adult friends sit across a folding table; person A quietly says, ‘I kept the ticket,’ and person B looks up without speaking. Use one restrained over-shoulder move. Preserve identities, wardrobe, eyelines, table objects, machine lights, and screen direction. No public-figure resemblance, subtitles, brand labels, extra people, or changing props.
Start Creating Source reserved
Reference generationImage to video16:95 sec

First-frame greenhouse portrait

Use the supplied rights-cleared image as the exact first frame and sole identity, wardrobe, proportion, palette, and environment reference. Animate the adult subject turning from a greenhouse worktable toward the camera while leaves move lightly in one consistent breeze. Use a slow five-second push-in. Preserve face, hair, navy jacket, hands, table objects, greenhouse geometry, and dawn light. No new person, source leakage, lip movement, logo, or identity drift.
Start Creating Source reserved
Reference generationProduct16:95 sec

Product reference orbit

Use the supplied authorized front, side, and rear images as exact references for one unbranded portable speaker. Create a restrained five-second tabletop move from front three-quarter to rear three-quarter view. Preserve every seam, edge radius, port, button, grille, material, color, and proportion; the references override stylistic invention. No new feature, logo, floating part, changing control layout, liquid, false scale, or text claim.
Start Creating Source reserved
Reference generationTransition16:95 sec

Locked weather transition

Use the supplied image as the exact first-frame composition of an empty coastal bus stop. Keep the camera locked while clear late-afternoon weather changes gradually to light rain at blue hour. Preserve shelter geometry, bench, road markings, cliff silhouette, horizon, lens, and camera position. Let clouds gather, grass move in one direction, and droplets appear on the glass. No people, lightning, flood, architecture change, camera motion, text, or logo.
Start Creating Source reserved
Reference generationCharacter4:35 sec

Character continuity close-up

Use the rights-cleared character image as the only identity, age, hairstyle, wardrobe, and proportion reference. Place the adult character in an archival reading room, opening one folder and noticing a missing page as the camera moves from medium profile to a close reaction. Preserve face, glasses, green cardigan, hands, folder, desk, lamp, and room geography. No real-person imitation, extra text, new accessory, duplicated hands, or identity change.
Start Creating Source reserved
Reference generationMotion16:95 sec

Motion-reference rehearsal

Use authorized video A only for the timing of one controlled turn, two steps, and a final held pose. Use authorized image B for the new original adult performer’s identity, silver rehearsal coat, and proportions. Replace the source setting with a plain dark studio and keep a locked full-body camera. Preserve image B throughout and do not copy the source face, clothing, background, camera, or audio. No unsafe motion, duplicated limb, costume redesign, or source leakage.
Start Creating Source reserved
Reference generationGames16:95 sec

Reference-led creature rise

Use the authorized concept image as the exact design reference for an original stone-and-moss guardian. Preserve silhouette, horn count, face, limb proportions, stone plates, moss placement, and scale. In a flooded underground temple, show the guardian rising slowly as water drains while a low wide camera retreats slightly. No franchise resemblance, extra limb, changing horn, rapid attack, HUD, text, copied setting, or redesign.
Start Creating Source reserved

How to use the Wan 2.0 archive

Treat the prompt as a production brief.

01

Write what must remain stable

State identity, product geometry, wardrobe, setting, screen direction, lighting, and source roles before describing motion. These invariants make drift visible during review.

02

Describe one observable change

Use a short ordered action and one motivated camera move. Avoid stacking unrelated transformations or relying on adjectives that do not change a visible production decision.

03

Keep the version boundary in the handoff

The prompts do not assert a Wan 2.0 specification. Verify the model label, accepted inputs, controls, limits, and terms in the current product before generation.

Version boundary checked August 15, 2026: current official Alibaba Cloud materials do not publish a Wan 2.0 specification in the listed model catalog.

Content context

Wan 2.0 prompts

Wan 2.0 prompt examples · Wan 2.0 video prompts · Wan image to video prompt

Evidence boundary — checked August 15, 2026: Alibaba Cloud’s current official catalog documents Wan versions beginning with 2.1 in the relevant legacy series. This page covers the Wan 2.0 search term without assigning later-version specifications, benchmark results, availability, pricing, speed, licensing, or output quality to it.

The Wan 2.0 label is treated as a version-verification query. Confirm the exact model shown in the current product surface before relying on any input, output, access, rights, safety, or commercial statement.

  • Video generation and editing — Alibaba Cloud Model Studio (2026-08-15): Current official Wan model catalog and versioned task families; the checked catalog begins its listed legacy series at Wan 2.1 rather than documenting a Wan 2.0 product specification.
  • Wan2.1: Open and Advanced Large-Scale Video Generative Models — Wan-Video (2026-08-15): Official Wan 2.1 repository identity, release history, open checkpoints, supported task families, license, and reproducible implementation notes.
  • Text-to-video image-to-video prompt guide — Alibaba Cloud Model Studio (2026-08-15): Current official Wan prompt structure for subjects, scenes, motion, camera, and negative constraints; version-specific examples are not attributed backward to Wan 2.0.

Reader notes

Before you generate.

How many prompts are in this Wan 2.0 collection?

There are 12 original SEELE Editorial directions: six text-led and six reference-led prompts across products, brands, ecommerce, games, motion, dialogue, characters, and transitions.

Can the prompts be used directly in SEELE?

Yes. Use in SEELE carries the complete prompt into Generation with the Wan template selected; verify the current Wan version and attach only authorized references.

Continue the work

Take a prepared brief into the workspace.

Continue with “Try it free” to confirm the current product context and next production decision.

Try it free