describe-image

To describe an image using a vision model, you provide local file paths or URLs. The system automatically spawns a subagent to process the inputs and return detailed descriptions covering scene, colors, composition, and mood.

Yes, you can analyze multiple images simultaneously. The system supports processing multiple image paths or URLs in a single request to generate detailed descriptions for each visual input.

Yes, image analysis works with both local file paths and URLs. You can provide either input type directly, and the vision model will process the visual data to extract scene and mood details.

Subagent utilization for image description automatically spawns a subagent equipped with a vision model when the current model lacks native vision support. This ensures advanced image processing without requiring manual model switching.

Generated image descriptions include detailed information about the scene, colors, composition, and mood. The vision models analyze these visual elements to provide a comprehensive textual understanding of the image content.