video-generation

SKILL.md

Video Generation: Create videos with Veo

Overview

Generate videos using Google's Veo models. Supports:

  • Text-to-video generation (4-8 seconds)
  • Image-to-video (animate a starting frame)
  • Native audio generation with dialogue and sound effects (Veo 3.1)

Reference: https://ai.google.dev/gemini-api/docs/video

How to use

bash ${CLAUDE_PLUGIN_ROOT}/scripts/gemini.sh --model=veo-3.1-generate-preview "YOUR VIDEO DESCRIPTION"

Arguments:

  • --model - Required: Use a Veo model (see Models below)
  • --file - Optional: Starting frame image for image-to-video
  • --aspect-ratio - Optional: 16:9 (default) or 9:16
  • --duration - Optional: 4, 6, or 8 seconds (default: 8)

Examples:

# Generate a video from text
npx -y superconductor-gemini-skills --model=veo-3.1-generate-preview "A serene mountain landscape at sunset, camera slowly panning right"

# Generate with specific aspect ratio and duration
npx -y superconductor-gemini-skills --model=veo-3.1-generate-preview --aspect-ratio=9:16 --duration=4 "A person walking through a busy city street"

# Include dialogue (Veo 3.1)
npx -y superconductor-gemini-skills --model=veo-3.1-generate-preview "Two friends at a coffee shop. One says \"Did you hear the news?\" The other replies \"No, what happened?\""

# Animate an image
npx -y superconductor-gemini-skills --model=veo-3.1-generate-preview --file=landscape.jpg "Camera slowly zooms in on the mountain peak"

# Fast generation (lower quality, quicker)
npx -y superconductor-gemini-skills --model=veo-3.1-fast-generate-preview "A cat playing with a ball of yarn"

Prompt tips

Effective prompts should include:

  • Subject: What appears in the video
  • Action: What the subject does
  • Style: Creative direction (cinematic, documentary, etc.)
  • Camera: Movement and positioning (pan, zoom, tracking shot)
  • Audio (Veo 3.1): Dialogue in quotes, sound effect descriptions

Output

Generated videos are saved to the current directory as gemini-video-{timestamp}.mp4.

Limitations

  • Video length: 4-8 seconds
  • Generation time: 11 seconds to 6 minutes
  • Videos stored on server for 2 days
  • All videos include SynthID watermarks

API Key

The GEMINI_API_KEY environment variable must be set. Get your key at: https://ai.google.dev/gemini-api/docs/api-key

Models

Model ID Features Speed
veo-3.1-generate-preview Native audio, best quality Normal
veo-3.1-fast-generate-preview Native audio, faster Fast
veo-3.0-generate-001 Audio support Normal
Weekly Installs
3
GitHub Stars
1
First Seen
Feb 9, 2026
Installed on
amp3
claude-code3
github-copilot3
codex3
kimi-cli3
gemini-cli3