Skip to main content

Google Veo 3.1

Google Veo 3.1 is ACT3 AI's primary high-quality video generation engine. It produces photorealistic, cinematic output from text prompts and is the best choice for final-quality renders where visual fidelity matters most. See cinematography for how to write effective camera instructions in your prompts.

What Google Veo 3.1 Does​

Veo 3.1 converts your shot prompts into short AI-generated video clips. Submit a detailed description of action, setting, and style, and Veo 3.1 returns a rendered clip in your Preview pane. Accepted clips go directly onto your timeline or into the Render Queue for final export.

Key Capabilities​

  • Prompt-to-Video — Text descriptions become cinematic video clips
  • Photorealistic Output — Film-quality visuals with natural lighting and motion
  • Resolution and Frame Rate — Up to 1080p at 30fps
  • Style Presets — Apply Google's built-in artistic filters or custom style parameters
  • Batch Requests — Queue multiple prompts for parallel generation and monitor in the Render Queue
  • Camera Instructions — Responds accurately to cinematography language: "dolly in," "tracking shot," "handheld," etc.

How to Use​

  1. In the Editor, select a shot or scene and click AI → Google Veo 3.1
  2. Your shot prompt is pre-populated from the shot description — edit as needed
  3. Include action, camera movement, lighting, and mood in the prompt
  4. Choose output settings: resolution and frame rate
  5. Click Generate — the request goes to Google Veo 3.1 and the clip appears in the Preview pane
  6. Accept the clip, refine the prompt and regenerate, or swap to a different engine

Prompt Tips for Veo 3.1​

Veo 3.1 performs best with clear, specific, cinematic prompts:

  • Keep prompts under 100 words
  • Use standard film terminology (dolly, pan, tracking, close-up, establishing shot)
  • Describe lighting explicitly (golden hour, film noir, studio lighting, overcast day)
  • Include mood and atmosphere words (tense, serene, chaotic, mysterious)
  • For longer takes, split into consecutive prompts and stitch in the Timeline

Example prompts:

  • "Medium shot of detective at a rain-streaked window, handheld camera, warm interior light vs cold blue exterior, reflections in glass, tense atmosphere."
  • "Wide establishing shot of Mars colony at sunrise, drone aerial, red dust plains, domed habitat, epic and lonely."
  • "Slow dolly in on two characters at a dinner table, candlelight, soft bokeh background, intimate and strained."

Credit Usage​

Google Veo 3.1 generation is billed per second of output:

  • Approximately 1 credit per second at 1080p
  • Draft-quality renders cost less
  • Actual rates are shown in the credit estimate before you confirm generation

Check your credit balance in the Credits Panel before queuing large batches.

Best Use Cases​

  • Final-quality video renders for distribution
  • Cinematic film and narrative content
  • Photorealistic environments and character scenes
  • Any content where visual realism is the priority

When to Use Other Engines​

  • Use Wan 2.1 for fast creative iteration and team collaboration previews
  • Use Grok Video for creative or stylistically distinctive output where photorealism is secondary

Troubleshooting​

Results don't match the prompt — Break the prompt into shorter, more specific phrases. Avoid conflicting directions in a single prompt.

Generation takes too long — Large batch requests may queue. Use Wan 2.1 for rapid iteration and switch to Veo 3.1 for final approved shots.

Credits running out — Use draft previews during the creative phase. Only commit Veo 3.1 renders to approved shots.