veo3-video-gen

What This Skill Does

The veo3-video-gen skill enables OpenClaw AI agents to generate high-quality video content directly through the Google Veo 3.1 API. Designed as a comprehensive CLI-integrated toolkit, it handles the end-to-end lifecycle of video production: from initial prompt-based generation and status polling to automatic downloading and advanced post-processing like video stitching. Whether you need short, punchy 8-second social media clips or longer, cohesive narrative sequences, this skill abstracts the complexities of API handshakes, media encoding, and duration management into a streamlined workflow.

Installation

To integrate this skill into your environment, ensure you have the uv package manager installed. Run the following command in your terminal:

clawhub install openclaw/skills/skills/bluelyw/veo3-video-gen

Ensure your GEMINI_API_KEY is configured in your environment variables. For multi-segment stitching, ensure ffmpeg is installed and available on your system path, as the script relies on it for seamless concatenation.

Use Cases

This skill is perfect for automated content creation workflows. Use it for:

• Social Media Marketing: Generate UGC-style clips, quick product reveals, or mood-setting b-roll for Instagram and TikTok campaigns.
• Rapid Prototyping: Create visual mood boards or storyboards to demonstrate a visual concept before committing to high-end production.
• Product Demos: Generate sequences demonstrating a product in different environments using the reference-image feature to maintain brand consistency.
• Iterative Content Generation: Build longer video assets by stitching together logically connected segments with style continuity.

Example Prompts

1. "Generate a 16-second cinematic video of a futuristic coffee machine brewing espresso, using the veo3-video-gen skill with 2 segments and a professional, clean lighting style."
2. "Create a 9:16 vertical video for a fashion brand showing a model walking through a rain-slicked city at night. Use the reference image at ./assets/model_style.jpg to guide the aesthetic."
3. "Stitch a 24-second story about a space explorer landing on a new planet. Generate 3 segments using the continuation style to ensure the narrative flows naturally between each clip."

Tips & Limitations

• Consistent Style: Always use the --base-style argument to prevent visual drift between segments.
• Resource Management: Veo requests are subject to quota limits. Monitor your GEMINI_API_KEY usage to avoid 429 errors.
• Hardware Dependency: ffmpeg is strictly required for any operation where --segments is greater than 1. Without it, the script will generate individual files but fail to combine them into your final video.
• Error Handling: If you receive a 503 error, the model is currently under high load; wait a few minutes before retrying your generation request.