Veo 3 Prompt Examples: 15 Templates for Cinematic AI Video

Google Veo 3 generates high-fidelity 1080p video with native audio sync, but output quality depends entirely on prompt precision. Unlike earlier models, Veo 3 interprets cinematographic language—lens specs, sensor characteristics, and directional audio cues—literally. Below are 15 battle-tested prompt structures you can adapt immediately, organized by production intent.

Core Prompt Architecture for Veo 3

Every effective Veo 3 prompt follows a consistent hierarchy: Subject + Action + Environment + Camera/Lens + Lighting/Color + Audio + Technical Modifiers. Omitting any layer forces the model to hallucinate defaults, often resulting in generic "AI look" footage. Veo 3's 4K-native latent space responds best when camera parameters mirror real-world optics.

• Subject/Action: Use present participle verbs ("walking," "gesturing") over static nouns.
• Camera: Specify focal length (mm), aperture (f-stop), and sensor format (Full Frame, Super 35).
• Lighting: Name the key light source ("5K HMI through 12x12 diffusion") and color temperature (Kelvin).
• Audio: Describe diegetic sounds with spatial cues ("footsteps crunching gravel, left-to-right pan").

Cinematic Narrative Templates

These prompts prioritize storytelling continuity and character presence. They work best for 8-second clips intended for editorial sequencing.

1. Character-Driven Close-Up

Prompt: Extreme close-up, elderly watchmaker's eye through 100mm macro lens f/2.8, dust motes dancing in shaft of 5600K daylight, shallow depth of field renders background workshop bokeh. Diegetic: faint tick of escapement mechanism, breath sounds. ARRI Alexa Mini LF texture, 23.976fps.

2. Environmental Reveal with Dolly

Prompt: Slow dolly push-in 30ft across abandoned server farm, 24mm T1.5 anamorphic, fluorescent 4100K flicker reflecting on raised floor tiles. Subject: lone technician in reflective vest typing, center frame. SteadyCam operator breathing audible. Cool cyan-teal grade, 16:9 safe area.

3. Handheld Verite Street Scene

Prompt: Handheld 35mm f/1.8, pushing through crowded Bangkok night market, ISO 3200 grain structure. Neon sign spill (magenta/cyan) on wet pavement. Subject: street vendor flipping roti, motion blur 1/48s shutter. Ambient: sizzle oil, Thai chatter, motorbike horns Doppler left-to-right. Documentary aesthetic.

Commercial & Product Visualization Prompts

Veo 3 excels at material rendering when surface properties are explicitly defined. These templates target product launch assets and spec spots.

4. Macro Product Hero Shot

Prompt: Rotating 360° hero shot luxury mechanical watch, 90mm shift lens f/11 for full depth, 5500K ring light + black flags for controlled reflections on sapphire crystal. Brushed titanium case, ceramic bezel indices catching specular highlights. Silent rotation motor whine. 4K ProRes 4444 alpha channel ready.

5. Liquid Dynamics Commercial

Prompt: High-speed 1000fps pour, single-origin espresso crema forming tiger stripes, 65mm macro f/16, dual Broncolor strobes 1/10,000s duration. Dark roast beans background, shallow focus transition to sharp liquid surface. Audio: viscous pour ASMR, cup clink. Slow-motion ramp to 24fps.

6. Automotive Exterior Golden Hour

Prompt: Low-angle 16mm fisheye, matte olive green Porsche 911 GT3 RS parked coastal cliff, 18:32 sunset 3200K warm rim light on carbon fiber wing. Wet sand reflections, incoming tide. Rolling shutter controlled. Engine idle rumble sub-bass 40Hz. Anamorphic flare streaks horizontal.

Atmospheric & Abstract Concepts

For mood reels, title sequences, and transition assets where narrative clarity is secondary to texture.

7. Volumetric Light Architecture

Prompt: God rays piercing gothic cathedral nave, 14mm rectilinear f/8, haze machine particulate density 40%. Dust particles tracked in 3D space, Mamiya RB67 medium format grain. Organ sustain C2 65Hz reverberant tail 4.2s. Exposure ramp +2EV over 8s.

8. Microscopic Cellular Division

Prompt: Darkfield microscopy 1000x, mitotic spindle fibers separating chromosomes, 560nm interference filter. Pseudocolor LUT: cyan/magenta phase contrast. No camera movement. Silence except subtle microscope stage hum 120Hz. Scientific visualization accuracy.

9. Procedural Landscape Flythrough

Prompt: Drone FPV 12mm f/2.8, diving through alien basalt column forest, bioluminescent moss 470nm emission. Unreal Engine 5 Nanite geometry density. Camera roll 360° at 0:04. Spatial audio: wind shear, resonant column vibration. 2.39:1 scope ratio.

Camera Movement & Technical Precision

Veo 3 distinguishes between mechanical camera moves (dolly, crane, technocrane) and optical moves (zoom, focus pull). Specify the mechanism.

10. Technocrane Rise with Tilt-Down

Prompt: Technocrane 50ft rise over 6s, simultaneous tilt-down 45°, 27mm T2.0 spherical. Subject: protest crowd becoming aerial pattern. Rolling shutter minimal. Audio: crowd roar fading to drone prop wash. ExpoDisc white balance lock.

11. Snap Zoom Crash Zoom

Prompt: 1970s zoom lens 25-250mm T3.5, crash zoom 25mm to 250mm in 12 frames, focus hunting momentarily. Grainy 16mm reversal stock emulation. Subject: pigeon taking flight. Audio: lens motor whine, wing flap transient. Freeze frame end.

12. Focus Pull Rack

Prompt: Rack focus 1.5m to infinity over 3s, 85mm f/1.2, foreground champagne flute to background couple kissing. Bokeh balls hexagonal (7-blade iris). Shallow DOF transition smooth. Audio: clink glass, muffled laughter sharpening. Canon K35 character.

Veo 3 vs. Competing Models: Prompt Sensitivity

Understanding how Veo 3's prompt adherence differs from Sora, Runway Gen-3, and Kling 1.6 prevents wasted iterations.

Veo 3 demands technical specificity; vague prompts yield technically perfect but creatively hollow footage. Sora tolerates poetic abstraction. Runway prefers style references over camera specs. Kling leads in character micro-expressions but struggles with complex camera rigs.

Negative Prompting Strategies

Veo 3's structured negative prompting accepts comma-separated exclusion tags. Append these to suppress common artifacts:

• Anatomy: extra fingers, fused digits, asymmetrical pupils, floating teeth
• Motion: morphing, sliding feet, temporal flicker, strobing, warp distortion
• Optical: lens breathing, chromatic aberration (unless specified), vignetting, sensor dust
• Audio: hallucinated speech, music bleed, static, clipping
• Style: cartoon, illustration, watermark, text overlay, UI elements

Example negative block: --neg "extra fingers, morphing, temporal flicker, lens breathing, hallucinated speech, cartoon, watermark"

Practical Workflow: From Concept to Final Clip

1. Define shot list with ASC standard notation (WS, MS, CU, ECU, OTS, POV).
2. Draft prompts using the hierarchy: Subject → Camera → Light → Audio.
3. Generate 4 variations per shot at 720p preview (faster iteration).
4. Select best take and upscale to 1080p with locked seed.
5. Extend duration using image-to-video on last frame if >8s needed.
6. Color grade in DaVinci Resolve using ACES 1.3 transform for consistency.
7. Audio sweetening: replace Veo 3's generated audio with Foley library for critical sync points.

Common Failure Modes & Fixes

• Floating objects: Add "weighted, grounded, contact shadows" to prompt.
• Temporal inconsistency: Specify "locked off tripod" or "motion control repeatable."
• Audio drift: Regenerate with "audio sync locked to timecode."
• Overexposure: Include "exposure compensation -1/3 stop, highlight protection."
• Wrong aspect ratio: Force "1.9:1 (1920x1080) letterbox safe."

Pro tip: Veo 3 respects EXIF-style metadata in prompts. Append "EXIF: 24mm f/2.8 1/48 ISO 800" to lock exposure triangle variables across generations.

Closing Takeaway

Veo 3 rewards cinematographic literacy. The model doesn't "imagine"—it simulates physics, optics, and acoustics based on your instructions. Treat prompts as technical directives, not creative writing. Master the 7-layer hierarchy, exploit negative prompting, and iterate at preview resolution. The gap between "AI video" and "virtual production" is prompt discipline.