Veo 3.1 Lite Image-to-Video: Turn Product Photos Into Clips in Under a Minute

Why Image-to-Video Changes the Workflow

Text-to-video is powerful, but it's probabilistic — you describe what you want and the model interprets it. Image-to-video is different: you provide the exact visual starting point, and the model animates from there.

For product work, this matters. Your product has a specific shape, color, material, and branding. Text prompts can't guarantee those details. An image can.

Veo 3.1 Lite supports image-to-video at 720p and 1080p, in both 16:9 and 9:16 formats, for 4s, 6s, or 8s durations. At 20 credits for 8 seconds on NanoBanana, it's cheap enough to run 5–10 variations on a single product shot and pick the best one.

How Image-to-Video Works in Veo 3.1 Lite

You provide:

1. One reference image — the first frame of the video
2. A text prompt — describes the motion, camera, and audio
3. Duration and aspect ratio — 4s/6s/8s, 16:9 or 9:16

The model generates video that starts from your image and animates outward. The image sets the visual identity; the prompt directs what happens next.

The key insight: the image handles "what it looks like," the prompt handles "what it does." Split the work that way and you get consistent, directed output.

What Makes a Good Input Image

Not all product photos work equally well. Here's what the model handles reliably versus what causes problems:

Best practice: Use the cleanest version of your product photo — the same one you'd use for an e-commerce listing. Remove backgrounds if possible. The cleaner the input, the more control you have over the output.

The First Frame / Last Frame Technique

Veo 3.1 Lite supports setting just the first frame (your product image as the opening shot). For controlled transitions — where you want the video to start at point A and end at point B — you can also set a last frame.

Use cases:

• Unboxing reveal: First frame = closed box. Last frame = open box with product visible.
• Before/after: First frame = problem state. Last frame = solved state.
• Rotate and settle: First frame = product at angle. Last frame = front-facing hero position.

This technique gives you cinematic control without complex prompting. The model interpolates the motion between your two anchors.

Prompt Templates by Use Case

These prompts are structured for image-to-video. The image provides the visual baseline — the prompt directs motion and feel.

Product: Hero Shot with Camera Move

The camera slowly dollies in toward the product.
Soft studio lighting, clean background.
No movement except the camera push.
SFX: silence.
Duration: 6 seconds.

The product rotates slowly 45 degrees clockwise, revealing its side profile.
Tabletop surface, warm side lighting catching texture details.
Camera static, 85mm lens.
SFX: subtle ambient studio hum.

Product: Lifestyle / In-Use

A hand reaches in from the right and picks up the product naturally.
Kitchen counter environment, warm afternoon light through a window.
Handheld camera feel, slight movement.
SFX: ambient kitchen sounds, soft handling noise.

The product is poured/opened/used in the natural way it's intended.
Close-up, 85mm. Soft natural light.
Focus shifts to the key moment of use.
SFX: the sound of the product being used.

Fashion / Apparel

Vertical 9:16 format.
The garment moves gently as if in a light breeze.
Model is still; only fabric has motion.
Outdoor natural light, overcast sky for diffused shadows.
SFX: wind, distant ambient sound.

Vertical 9:16 format.
A close-up of the fabric texture. Camera pulls back slowly to reveal the full garment.
Shallow depth of field, 85mm.
SFX: silence.

Food & Beverage

Steam rises gently from the dish/drink.
Overhead camera, static.
Warm practical lighting, dark background for contrast.
SFX: ambient café or kitchen sound, very low.

Close-up. The liquid pours slowly into frame from above, filling the glass.
Camera static, 85mm. Black background, single side light.
SFX: the sound of liquid pouring, ice clinking.

Social Hook (Vertical, 0–4 seconds)

Vertical 9:16. Close-up.
The product spins once and comes to a stop facing the camera.
Bright, clean background. Quick, energetic motion.
SFX: a short whoosh sound as it spins, then stops.
Duration: 4 seconds.

Vertical 9:16. Medium shot.
The product drops into frame from above and lands with a satisfying impact.
High-contrast background. Slight slow-motion on the impact.
SFX: a clean thud as it lands.
Duration: 4 seconds.

Full Workflow: Photo to Posted Video

Common Issues and Fixes

The product looks distorted after a second or two The model is over-animating. Reduce motion in your prompt: add camera static or minimal movement, only [specific element] moves.

Background changes unexpectedly Your background has too much detail and the model is reinterpreting it. Reshoot on a simpler background, or add background unchanged, only product moves to your prompt.

The video looks like a slideshow, not smooth motion Prompt for continuous motion: smooth continuous camera move or fluid 360 rotation. Avoid start/stop action descriptions.

Portrait image shows black bars in 9:16 output Crop or pad your input image to 9:16 before uploading. Mismatched aspect ratios cause the model to letterbox.

What Veo 3.1 Lite Can't Do (For Product Work)

• No 4K — max is 1080p. Fine for web and social; not suitable for large-format print or digital signage.
• No Extension — you can't extend a generated clip beyond 8 seconds in Lite tier.
• No multi-product comparison — animating two products interacting is unreliable. Generate separately and edit together.
• No text overlay — don't rely on the model to add readable text/prices/callouts. Add those in post.

Try It: Free First Generation

NanoBanana's Veo 3.1 Lite generator supports image-to-video with the same prompt interface. Upload your product photo, paste a prompt from above, pick your format, and generate.

→ Try Veo 3.1 Lite Image-to-Video

20 credits for 8 seconds. Half the cost of Veo 3.1.

FAQ

Disclosure

Video examples use footage from the Veo 3.1 model family. Workflow recommendations are based on practical testing of image-to-video generation. Results vary by input image quality and prompt specificity.