Model · Google DeepMind

Reference-to-video, Veo 3.1 consistency.

Google DeepMind's Veo 3.1 — the strongest reference-to-video model available. Lock character identity across multiple shots with built-in audio support and flexible aspect ratios.

8sDefault duration

120sMax chained

4.5Credits/sec

Generate with Veo 3.1 Browse apps

◆ Model

Veo 3.1

Model · Google DeepMind

Capabilities

What it does.

06Features

Multi-reference

Feed multiple character references to lock identity across shots.

Image-to-video

Animate any still with reference-conditioned motion fidelity.

Audio-on / off

Choose audio synthesis on the fly — saves credits when silent.

Auto-fix prompts

Server-side prompt rewriting if content policy trips a flag.

Resolution control

720p or 1080p, 16:9 / 9:16 / 1:1 — render where it'll be watched.

Dual provider

Available via both Fal.ai and Kie.ai accounts — automatic load balancing.

Samples

Generated here.

Every clip below was rendered through Prototipal — no curation, no cherry-picked external content.

06Samples

Character consistency

Reference-to-video

Multi-shot scene

Audio-on talking head

Cinematic landscape

Product showcase

Related Apps

Use it across the catalog.

— 4.5 credits/sec audio-on · 3.0 credits/sec audio-off — pay only for what you render. —

Reference-grade video at production pricing.

8-second clip with audio: 36 credits

Generate with Veo 3.1