gemini-video-analyzer

What This Skill Does

The gemini-video-analyzer is a powerful OpenClaw AI agent skill designed to leverage Google's Gemini multimodal capabilities for direct video analysis. Unlike traditional methods that rely on labor-intensive frame extraction or pre-processing, this skill enables the agent to process video files natively. By analyzing video content at 1 frame per second, the skill maintains temporal context, allowing it to understand motion, transitions, and audio-visual cues simultaneously. Whether you are dealing with screen recordings, instructional tutorials, or long-form meetings, this tool turns raw video data into structured, actionable insights without the need for manual transcription or visual review.

Installation

To integrate this skill into your environment, use the OpenClaw command-line interface. Ensure you have your GOOGLE_AI_API_KEY ready, as it is required for authentication with Google's Files API. Run the following command in your terminal:

clawhub install openclaw/skills/skills/aiwithabidi/a6-gemini-video-analyzer

Once installed, configure your environment by adding your API key to your .env file to allow the scripts to communicate with the Gemini models seamlessly.

Use Cases

This skill is highly versatile and serves various professional and personal needs. Use it for:

• Technical Documentation: Automatically extract text and UI elements from software walkthroughs to create written guides.
• Quality Assurance: Submit screen recordings of bugs to let the AI describe exactly what went wrong and identify failure points.
• Content Summarization: Process long meeting recordings to generate concise minutes, key discussion points, and action items.
• Media Comparison: Analyze multiple video files simultaneously to highlight differences, trends, or stylistic variations.

Example Prompts

1. "Analyze this screen recording and extract all the error messages shown in the terminal window."
2. "Summarize the key takeaways from this hour-long tutorial video and list the steps in bullet points."
3. "Compare these two videos and tell me which one displays the more efficient workflow for setting up a database."

Tips & Limitations

• File Limits: Ensure your video files are under 2GB. Supported formats include MP4, MOV, WebM, and more.
• Storage: Files uploaded via the Gemini Files API are temporary and will be automatically purged after 48 hours for your privacy.
• Performance: Use gemini-2.5-flash for most tasks to benefit from speed and cost-efficiency. If you require deep reasoning for complex visual tasks, use the --model gemini-2.5-pro flag.
• Audio: The model excels at understanding audio context, so ensure your source videos have clear audio tracks for the best results.