Generate or edit images from text prompts or reference images with customizable sizes using Agnes Image 2.1 Flash model and receive direct image URLs.
# Agnes Image Generator A skill for generating and editing images using Agnes Image 2.1 Flash model. ## Capabilities - Text-to-image generation - Image-to-image editing (with URL or Data URI input) - Returns direct image URL for easy access - Supports custom sizes (default 1024x1024) ## Usage When the user says "生成图片" or any image generation request, this skill will be triggered automatically if configured as the primary image generation tool. The skill accepts: - `prompt` (string, required): Text description of the image to generate - `image` (string, optional): URL or Data URI of reference image for img2img - `size` (string, optional): Output size, e.g., "1024x1024" (default) ## Implementation - Calls Agnes API endpoint: `https://apihub.agnes-ai.com/v1/images/generations` - Uses Bearer token authentication from `AGNES_API_KEY` environment variable - Requests `response_format: "url"` for direct accessible links - For img2img, passes image URL in `extra_body.image` array ## Configuration Set your Agnes API key: ```bash export AGNES_API_KEY="your-api-key-here" ``` Or configure in OpenClaw auth profile (if supported). ## Notes - Model: `agnes-image-2.1-flash` - Pricing: $0.003 per image (as per Agnes documentation) - Supports high information density images and composition preservation for img2img.
don't have the plugin yet? install it then click "run inline in claude" again.