# Veo 3.1

Veo 3.1 in Sloosh: Google's 8-second video with sound, up to 4K, with 3 references or start and end frames.

Last updated: 2026-10-05

**Key facts**

- Veo 3.1 is Google's high-fidelity video model, and every Veo 3.1 clip in Sloosh is 8 seconds long with sound.
- Veo 3.1 outputs 720p, 1080p or 4K in Sloosh, in 16:9 or 9:16 only.
- Veo 3.1 takes up to 3 reference images, or a start frame with an optional end frame, but not both in one clip.
- Veo 3.1 has a seed setting in Sloosh to repeat a result, and accepts prompts of up to 3,000 characters.

Web page: https://sloosh.ai/docs/veo-3-1

Veo 3.1 is Google's high-fidelity video model. Every clip is 8 seconds in Sloosh and comes with sound: dialogue, effects and ambience. It can output up to 4K.

## When to use it

- One polished, cinematic shot with sound.
- A start and end frame with a described move between them.
- Only 16:9 or 9:16, and a start frame can't be combined with reference images.

## How to prompt it

- Google's formula: cinematography, subject, action, context, then style and ambience.
- Use camera words: dolly, tracking, crane, aerial, slow pan, close-up, low angle, shallow depth of field.
- Dialogue goes in quotes; sound effects start with "SFX:"; ambience with "Ambient noise:".
- There are no reference tags: describe each reference subject and Veo matches them.
- Describe what you want rather than writing "no".

## Example

Text only: one shot, with the sound written into the prompt.

[![A hand tapping a black wireless speaker on a walnut sideboard in a sunlit loft](https://sloosh.ai/docs/m-veo-3-1.webp)](https://sloosh.ai/docs/m-veo-3-1.mp4) (video)

_16:9, 720p, 8 seconds, Enhance off. Clips here play without sound; press one to pause._

**Prompt**

```text
Slow dolly-in, close-up: a matte black wireless speaker sits on a walnut sideboard in a sunlit loft apartment. A hand taps its top; a ring of soft white light pulses and the room fills with warm, upbeat acoustic guitar. Dust motes drift through golden late-afternoon light, shallow depth of field, cosy Scandinavian interior. SFX: a soft tap, then the gentle hum of the city through an open window. Clean, premium commercial look.
```

## Settings in Sloosh

- Aspect ratio: 16:9 or 9:16.
- Resolution: 720p, 1080p or 4K.
- Length: 8 seconds.
- References: up to 3 images.
- Start and end frames: both optional (an end frame needs a start frame).
- A start or end frame can't be combined with reference images.
- Sound: always on.
- Seed: available, to repeat a result.
- Prompt: up to 3,000 characters.

Open it in [Video](https://sloosh.ai/studio/video) and pick Veo 3.1 in the model chip. The Generate button shows the cost first.

## Frequently asked questions

### Can Veo 3.1 make a video longer than 8 seconds in Sloosh?

No. Every Veo 3.1 clip in Sloosh is 8 seconds. For a longer single clip, [Seedance 2.5](https://sloosh.ai/docs/seedance-2-5.md) runs up to 30 seconds; or make several Veo clips and join them with Video Concat in [Spacelab](https://sloosh.ai/docs/spacelab.md).

### How do I add dialogue and sound effects in Veo 3.1?

Write dialogue in quotes, start sound effects with "SFX:" and ambience with "Ambient noise:", for example "SFX: thunder cracks in the distance". Veo 3.1 always makes sound, so without these lines it guesses the audio.

### How do I refer to reference images in a Veo 3.1 prompt?

Veo 3.1 uses no tags like "Image 1": describe each subject and Veo matches the reference by what it shows. Google's example reads: "Using the provided images for the detective, the woman, and the office setting, create a medium shot of ...".

### Can Veo 3.1 use an end frame without a start frame?

No. In Sloosh a Veo 3.1 end frame needs a start frame. With both, describe the camera move between them, such as a smooth 180-degree arc from the first view to the last, and leave out reference images.

### Do Veo 3.1 videos have a watermark?

Google marks every Veo 3.1 output with SynthID, an invisible watermark. Google also notes that Veo 3.1 has been evaluated in English only, so write prompts in English for the most predictable results.

## Sources

- [Gemini API: generate videos with Veo 3.1](https://ai.google.dev/gemini-api/docs/veo)
- [Google Cloud: prompting guide for Veo 3.1](https://cloud.google.com/blog/products/ai-machine-learning/ultimate-prompting-guide-for-veo-3-1)