Wan 3.0 Prompt Guide — Write Prompts That Work

Wan 3.0 prompts control an entire video production pipeline — camera movement, character appearance, scene lighting, audio mix, and multi-shot structure — all from a single text input. Unlike image prompts or standard chatbot instructions, a well-structured Wan 3.0 prompt can reference up to 12 uploaded assets using @reference syntax, and trigger automated multi-shot editing via AI Director mode.

The anatomy of a complete prompt follows this five-segment structure:

[Scene + Lighting] -> Camera [Movement] -> [Subject + Appearance] -> [Audio Tone] -> @Image1 @Audio1

Every strong Wan 3.0 prompt maps to one of three formulas. Pick the right formula for your mode before writing a single word — it determines what Wan 3.0 optimizes for.

[Shot Type] of [Subject + Appearance], [Action], [Setting + Lighting]. Camera [Movement]. [Audio description]. [Style note]. 4K.

Medium tracking shot of a man in a navy wool coat standing at a fog-covered harbor pier at sunrise, gazing toward open water. Camera slowly cranes upward to reveal the full skyline. Low distant foghorns, gentle waves, orchestral score swells softly. Cinematic, 35mm film grain. 4K.

Shot 1 [0–5s]: [Shot type] — [Scene content].
Shot 2 [5–12s]: [Shot type] — [Scene content].
Shot 3 [12–20s]: [Shot type] — [Scene content].
Shot 4 [20–28s]: [Shot type] — [Scene content].
Overall tone: [audio and mood description].
Character reference: @Image1.

[Scene description]. Character appearance: @Image1. Camera style reference: @Video1. Background music tone: @Audio1. [Style and resolution notes].

Camera language is the single biggest lever in Wan 3.0 output quality. Prompts with explicit shot type and named camera movement consistently outperform generic descriptions. Use the vocabulary below exactly — Wan 3.0 is trained on these terms.

Wide shot · Medium shot · Close-up · Extreme close-up · Over-the-shoulder shot · POV shot · Aerial shot · Dutch angle · Two-shot · Insert shot

- Slow push in          -> camera moves toward subject
- Tracking shot         -> camera moves alongside moving subject
- Dolly zoom            -> zoom + physical move, creates vertigo
- Crane up              -> camera rises vertically, reveals skyline or context
- Crane down            -> camera descends, closes in on subject
- Orbit                 -> camera circles subject (specify clockwise/counterclockwise)
- Handheld              -> intentional shake, naturalistic, documentary feel
- Static locked         -> no movement, deliberate stillness
- Whip pan              -> fast horizontal snap between subjects
- Dutch tilt            -> angled frame, psychological unease
- Rack focus            -> shifts focus between foreground and background

Extreme close-up of coffee being poured into a white ceramic cup on a marble countertop, morning light through frosted glass. Camera slowly orbits counterclockwise while rack focusing from steam to the cup surface. Rich coffee pour sounds, quiet kitchen ambience, gentle jazz in background. Warm golden color grade. 4K.

AI Director mode lets you specify up to 6 independent shots in a single generation pass. Wan 3.0 handles framing, transitions, and cross-shot character and environment consistency automatically — you define what happens in each shot, not how the cuts work.

Rules for AI Director Prompts

Four things to include in every AI Director prompt:

1. Shot number and time range for each shot [Shot 1 [0–6s]]
2. Shot type for each shot (Wide, Medium, Close-up, etc.)
3. A brief scene description per shot — one to two sentences max
4. An overall tone line at the end covering audio and mood

Full AI Director Example — Product Launch

Shot 1 [0–6s]: Aerial shot — overhead view of a minimalist product on white marble, soft studio light, camera slowly descends.
Shot 2 [6–12s]: Medium shot — hands carefully unbox the product, crisp packaging sounds, warm studio light.
Shot 3 [12–18s]: Close-up — product features highlighted with subtle camera orbit, macro lens detail.
Shot 4 [18–24s]: Over-the-shoulder shot — person using the product naturally in a home setting, golden hour light.
Shot 5 [24–30s]: Wide shot — product placed on table, camera pulls back slowly to reveal lifestyle context, fade to brand color.
Overall tone: minimal ambient sound design, no dialogue, understated orchestral swell at Shot 5. Color grade: clean, warm whites. 4K.
Character reference: @Image1.

Identity Lock saves a character's visual profile after the first generation. Referencing that profile in later sessions produces the same character in new scenes without re-describing appearance. This is Wan 3.0's answer to one of the biggest pain points in AI video: character drift across clips.

First Generation — Establishing the Character

On the first generation, describe the character in full physical detail. The more specific, the stronger the Identity Lock profile.

Medium shot of a woman, early 30s, East Asian features, straight shoulder-length black hair, warm skin tone, wearing a structured white blazer over a cream silk top. She stands at a glass-walled office, afternoon city light behind her. Camera holds static. Audio: quiet office ambience, no dialogue. 4K. [Save as: brand-spokesperson-01]

Subsequent Generations — Calling the Profile

Character: @brand-spokesperson-01. Wide shot of the same character walking through a sunlit park, casual weekend outfit — dark jeans, white linen shirt. Camera tracks alongside her. Audio: birdsong, light acoustic guitar, no dialogue. 4K.

Wan 3.0 generates audio — dialogue, ambient sound, effects, and background score — in the same pass as the video. No post-production audio sync is required. Your prompt controls every layer of the audio output separately.

Audio Prompt Structure

Audio: [Dialogue description or "no dialogue"]. [Ambient environment sound]. [Sound effects]. [Score/music description]. [Room tone or acoustic space].

Audio Prompt Examples by Use Case

- Product reveal:
  Audio: no dialogue. Studio silence. Subtle product design sound (click, texture). Cinematic bass hum. Luxury brand aesthetic.

- Spokesperson (lip sync from @Audio1):
  Audio: phoneme-accurate lip sync to @Audio1. Natural room reverb, softbox lighting acoustic. No background music.

- Short film scene:
  Audio: sparse dialogue between two characters — tense, quiet. Interior ambient hum. No score until Shot 4.

- Social content:
  Audio: trending upbeat lo-fi instrumental. No dialogue. Satisfying product interaction sounds at key moments.

- Multilingual ad (12 languages supported):
  Audio: lip sync to @Audio1 (Spanish voiceover). Phoneme-accurate. Neutral room tone. No music.

Adjust the bracketed fields and generate. Every template is structured for the right formula and audio syntax.

Slow-motion product reveal of a [product name] rotating on a white marble pedestal, soft directional studio lighting, subtle smoke rising from below. Camera orbits clockwise at 45-degree angle. Close-up on key design details at 15s mark. Audio: deep cinematic bass hum, subtle product texture sound, no dialogue. Luxury brand aesthetic, desaturated background, product in full saturation. 4K.

Vertical 9:16 frame. Close-up of [subject action] in [setting], natural light. Handheld energy, dynamic. Quick cuts implied by subject movement. Audio: trending upbeat lo-fi music, no dialogue, satisfying sound effect at key moment. Text overlay space at bottom third. Warm vivid color grade. 4K. 15 seconds.

Shot 1 [0–7s]: Wide establishing — [location], [time of day], no characters yet.
Shot 2 [7–15s]: Medium — [character description] enters frame, [action].
Shot 3 [15–22s]: Close-up — [emotional beat, object detail, or reaction].
Shot 4 [22–28s]: Wide pullback — [resolution or open ending].
Overall tone: sparse [genre] score, minimal dialogue, naturalistic ambient sound. 4K.
Character: @Image1.

Product reference: @Image1. The product is unboxed on a clean white surface, hands gently lift it from clean packaging. Camera slowly pushes in as the product is held up for inspection in bright studio light. Audio: satisfying unboxing sounds — tissue paper, crisp packaging material, subtle product weight sounds. No music, no dialogue. 4K.

Shot 1 [0–6s]: Aerial — [company HQ or city skyline], sunrise, camera descends slowly.
Shot 2 [6–14s]: Medium — diverse team collaborating around a table, natural office light, no posed expressions.
Shot 3 [14–20s]: Close-up — hands at keyboard or on product, focused detail work.
Shot 4 [20–26s]: Wide — open-plan office, energy and motion.
Shot 5 [26–30s]: Wide pullback — group moment, faces visible, natural smiles.
Brand color reference: @Image1 (logo). Audio: corporate orchestral lift, subtle, no dialogue. 4K.

Character appearance: @Image1. Medium close-up — character speaks directly to camera in a modern, softly lit home studio setting. Phoneme-accurate lip sync to @Audio1 ([language] voiceover). Audio: natural room tone, slight reverb, no background music. Clean neutral background, no distractions. 4K. Suitable for multilingual ad campaign.

Medium tracking shot of an athlete in [sport/activity], outdoor setting, [time of day] light. Camera tracks alongside at matched speed. High-energy handheld feel with intentional motion blur on fast movement. Audio: fast rhythmic music — no lyrics — rising intensity, footstep sounds synced to movement, no dialogue. High contrast color grade, vivid colors. 4K. 20 seconds.

Extreme close-up of [food or drink] being prepared on [surface], natural window light from left. Slow motion at 60fps. Camera holds static then slow push in at 6s. Audio: satisfying preparation sounds — sizzle, pour, texture — no music, no dialogue. Warm golden color grade, shallow depth of field. 4K. 12 seconds.

The most common Wan 3.0 prompt failures share a pattern: too vague, leading with adjectives, or missing camera and audio direction entirely. Here are six common mistakes with corrected versions.

Beautiful cinematic video of a woman walking.

Medium tracking shot of a woman in a cream linen dress walking along a coastal cliff path at golden hour. Camera tracks alongside at walking pace. 4K.

Close-up of coffee being poured. Morning light. Warm.

Extreme close-up of espresso being poured into a white ceramic cup on a marble surface, morning window light. Camera holds static then slowly pushes in at 5s. Audio: rich pour sounds, quiet kitchen ambience. 4K.

A woman running through a forest at dawn. 4K.

Medium tracking shot of a woman trail running through a dense green forest at dawn, mist rising. Camera tracks alongside at running pace, handheld energy. Audio: rhythmic footsteps on earth, birdsong, gentle instrumental score rising — no dialogue. 4K.

Shot 1: Wide shot of city. Cut to Shot 2: Medium shot of character entering. Then close-up of their face.

Shot 1 [0–6s]: Wide — city skyline at dusk.
Shot 2 [6–14s]: Medium — character enters building lobby, confident stride.
Shot 3 [14–20s]: Close-up — face in elevator reflection, determined expression.

Copy these terms directly into your prompts. Wan 3.0 maps each to specific generation behavior.