🎬 Veo — Video Generation Skill for OpenClaw

Generate videos with Google's Veo via the Gemini API. Default model is Veo 3.1 Fast — great quality with fast turnaround. Switch to quality for maximum output.

Features

• Text-to-video — generate from a prompt
• Image-to-video — animate from a starting image
• Element references — provide assets (objects, characters, scenes) to appear in the video
• Style references — provide aesthetic references (lighting, texture, color palette)
• Model aliases — fast, quality, 3.1, 3.0, 2.0 for easy switching
• 16:9 / 9:16 aspect ratios, 5–8 second duration

Install

clawhub install veo

Requirements: uv

curl -LsSf https://astral.sh/uv/install.sh | sh

Quick Start

uv run scripts/generate_video.py \
  --prompt "A golden sunset over the ocean, waves gently crashing" \
  --filename "sunset.mp4" \
  --api-key YOUR_GEMINI_API_KEY

Models

Alias	Model	Notes
fast (default)	veo-3.1-fast-generate-preview	Fastest, great quality
quality	veo-3.1-generate-preview	Highest quality, slower
3.0	veo-3.0-generate-001	Veo 3.0 quality
3.0fast	veo-3.0-fast-generate-001	Veo 3.0 fast
2.0	veo-2.0-generate-001	Most compatible

# Fast (default)
--model fast

# Highest quality
--model quality

# Or use full model name
--model veo-3.1-generate-preview

# List all available models
--list-models

Examples

Text-to-video

uv run scripts/generate_video.py \
  --prompt "A golden sunset over the ocean" \
  --filename "sunset.mp4"

High quality

uv run scripts/generate_video.py \
  --prompt "A golden sunset over the ocean" \
  --model quality \
  --filename "sunset-hq.mp4"

Image-to-video

uv run scripts/generate_video.py \
  --prompt "The character slowly turns to face the camera" \
  --input-image character.png \
  --filename "character-turn.mp4"

With element + style references

uv run scripts/generate_video.py \
  --prompt "The red vase sits on the table" \
  --element red-vase.png \
  --style painterly-style.png \
  --filename "vase.mp4"

Vertical video (social)

uv run scripts/generate_video.py \
  --prompt "Close up of coffee being poured, slow motion" \
  --aspect-ratio 9:16 \
  --filename "coffee-pour.mp4"

Options

Flag	Default	Description
--prompt	required	Text description
--filename	required	Output path (.mp4)
--model	fast	Model alias or full name
--duration	8	Seconds (5–8)
--aspect-ratio	16:9	16:9 or 9:16
--negative-prompt	—	What to avoid
--seed	—	Reproducibility seed
--count	1	Number of videos
--input-image	—	Starting frame (image-to-video)
--element	—	Asset reference image (repeatable)
--style	—	Style reference image (repeatable)
--api-key	—	Gemini API key (or GEMINI_API_KEY env)
--list-models	—	List available Veo models and exit

Gemini API Limitations

These features require Vertex AI and are not available on the standard Gemini API:

• --last-frame — workaround: generate two clips and stitch with ffmpeg
• --resolution — API controls this automatically
• --fps / --generate-audio

API Key

Get a Gemini API key at aistudio.google.com. Pass via --api-key or set GEMINI_API_KEY.

Notes

• Generation takes 2–5 minutes (async polling)
• --count > 1 saves as name-1.mp4, name-2.mp4, etc.
• Videos download from a temporary URI after generation

License

MIT