How to use Veo 3 to make cinematic AI videos with sound

Veo 3 is Google DeepMind’s AI video model, and Veo 3.1 is the latest model in the lineup. It turns a text prompt into an 8-second clip with sound built in. The easiest way to use Veo 3 is on Picsart, which runs the latest Veo 3.1 with no install and no API setup: open the Picsart AI Video Generator, select Veo 3.1, describe the shot you want, and generate.

Veo 3 stands out for one reason above all: it generates video and synchronized audio, including dialogue, sound effects, and music, in a single render. This guide covers what Veo 3 is, what the latest Veo 3.1 adds, how to make your first clip step by step, how to write Veo 3 prompts, and the mistakes to avoid. By the end, you will have a finished, sound-on video in a few minutes.

What is Veo 3?

Veo 3 is Google DeepMind’s cinematic AI video model, and Veo 3.1 is the newest model in the lineup, the version this guide focuses on. It generates 8-second videos at 720p, 1080p, or 4K, with natively generated audio in the same pass. Where most video models hand back a silent clip, Veo 3 delivers the scene and its soundtrack together.

The main capabilities:

• Text to video with sound. Describe a scene and Veo 3.1 renders it with dialogue, sound effects, and music that match the action.
• Resolution up to 4K. Choose 720p, 1080p, or 4K. Higher resolution takes longer to render.
• Landscape or portrait. Generate 16:9 for widescreen or 9:16 for vertical social video.
• Image to video. Start from a still image and let Veo 3.1 bring it to life.
• Up to 3 reference images. Provide images of a person, character, or product to keep the subject consistent across the video.
• First and last frame. Set a starting frame and an ending frame, and Veo 3.1 fills in the motion between them.
• Video extension. Continue a clip you generated earlier to build a longer sequence.

On Picsart, Veo 3.1 runs in the AI Video Generator and AI Playground alongside other leading video models, so you can compare it against Kling, Runway, Seedance, and more on the same prompt.

Veo 3.1 at a glance

What changed from Veo 3 to Veo 3.1

Veo 3.1 keeps everything that made Veo 3 stand out, the synchronized audio and up-to-4K output, and adds four things on top:

• Portrait video. Native 9:16 output, so a clip is ready for Reels, Shorts, and TikTok without cropping.
• Reference images. Feed in up to 3 images of a person, character, or product to keep the subject looking the same from start to finish.
• First and last frame. Define the opening and closing frames, and Veo 3.1 animates the transition between them.
• Video extension. Continue an existing clip to push past the 8-second mark and build a longer sequence.

The short version: Veo 3.1 gives you more control over consistency and format. Since it is the current version, it is the one to use, and it is what Picsart runs.

How to use Veo 3 on Picsart, step by step

You can make a Veo 3.1 clip in a few minutes, with no account juggling or setup. Here is the full flow:

How to use Veo 3 on Picsart, step by step

1. Open the AI Video Generator

Open the Picsart AI Video Generator or AI Playground.

Open the AI Video Generator

2. Select Veo 3.1

Select Veo 3.1 as the model.

3. Write your prompt

Write a prompt describing the scene, the action, the camera move, and any dialogue or sound.

4. Add a reference image or frame

Add a reference image or a starting frame if you want to guide the subject or the opening shot.

5. Set aspect ratio and resolution

Set the aspect ratio and resolution, 16:9 or 9:16, up to 4K.

6. Generate and refine

Generate, preview the clip with sound, then refine the prompt or download.

Everything runs on one credit balance, and it is free to try, so you can experiment before committing.

Generate your first Veo 3 video: a worked example

The fastest way to learn Veo 3 is to make something. Say you want a food clip. Start with a prompt that names the scene, the action, and the sound:

“A close-up of a black skillet on a stove, colorful stir-fry vegetables, red and yellow bell peppers and snap peas, tossing and sizzling with rising steam, warm kitchen light from a window behind. Gentle sizzle and upbeat background music.”

Select Veo 3.1, set 16:9 at 1080p, and generate. In a few minutes you get an 8-second clip of the vegetables sizzling and steaming, with the sound already mixed in. From there, refine in small steps. Want it for social? Switch to 9:16 and regenerate. Want a voice? Add a line in quotes, such as a chef saying “Almost ready.” Want the same dish in a second shot? Add a reference image so it stays consistent. Each pass gets you closer without starting over.

How to write Veo 3 prompts

Veo 3 prompts reward detail. The model follows a well-structured description closely, so the more you specify, the more control you get. A strong prompt covers five things:

• Scene and subject. What and who is in the shot, and the setting.
• Action. What happens across the 8 seconds.
• Camera. The shot type and movement, such as a slow push-in, a drone pull-back, or a handheld close-up.
• Style and lighting. Cinematic, stop-motion, golden-hour, neon night, and so on.
• Audio. The dialogue, sound effects, or music you want, since Veo 3.1 generates sound too.

Veo 3 handles three prompt styles especially well. Here is an example of each:

• Dialogue and sound effects: “A close-up of a weathered older man sitting alone in a dim, stone-walled room lit by a single wall lamp. He looks down and murmurs, ‘It was here all along.’ Quiet, warm ambience with a faint crackle.”
• Cinematic realism: “A low tracking shot of a lone surfer paddling out at dawn, mist hanging over the water, the first golden light catching the waves. Gentle ocean swell and distant gulls.”
• Creative animation: “A charming paper-craft animation of a tiny sailboat drifting across a folded-paper ocean, cotton-ball clouds overhead, a lighthouse blinking in the distance, soft whimsical music.”

For dialogue, write the line inside quotes so Veo 3.1 renders it as spoken audio. For mood, name the music or ambience directly. And because a clip is only 8 seconds, keep the action to a single clear beat rather than a whole story. For a full library of ready-to-use ideas, see these Veo prompts.

What you can create with Veo 3

Veo 3.1 is built for finished, sound-on video, which opens up a lot:

• Cinematic clips. Establishing shots, product hero moments, and moody scenes with a full soundtrack.
• Talking videos. Characters that speak, with dialogue rendered as synchronized audio.
• Social video. Native 9:16 clips for Reels, Shorts, and TikTok, sound included.
• Ads and promos. Short product spots with voiceover, sound effects, and music in one render.
• Animation. Stop-motion, 3D, and stylized looks from a single descriptive prompt.
• Image to video. Turn a still, a product shot, or an illustration into a moving clip.

Common Veo 3 prompt mistakes to avoid

A few habits trip people up when they start:

• Cramming a whole story into 8 seconds. Pick one clear beat. A single moment lands; a full narrative gets rushed.
• Forgetting the audio. Sound is half of what Veo 3.1 makes, so name the dialogue, sound effects, and music instead of leaving it to chance.
• Vague camera direction. “A nice shot” leaves too much open. Name the shot type and the movement.
• Expecting consistency without references. If a face or product must stay the same, add up to 3 reference images rather than re-describing it.
• Rendering everything at 4K. Draft at 720p or 1080p to iterate fast, then render the keeper at 4K.

Tips for the best results

A few habits get more out of every generation:

Start creating with Veo 3.1 on Picsart

Using Veo 3.1 is as simple as describing a shot, then guiding it with reference images, aspect ratio, and resolution. Open the Picsart AI Video Generator, select Veo 3.1, and make your first sound-on clip in a few minutes.