@apresai/gimage-mcp

Generate AI images using multiple providers (Gemini, Vertex AI, xAI Grok). Call list_models first to see available providers with pricing and capability flags. Quick start: generate_image(prompt='sunset', output='~/Desktop/sunset.png') uses the default Gemini provider. The 'count' parameter requests multiple images (Grok returns N exactly; the Gemini family is best-effort), saved as image_1.png, image_2.png, etc. Styles, seeds, image_size, and other options depend on provider capabilities; options the chosen provider does not support are reported in the response 'warning' field and ignored. IMPORTANT: Always specify output path (e.g., ~/Desktop/image.png).

Resize an image to specific dimensions using high-quality Lanczos resampling. Supports three modes: 'crop' (default) preserves aspect ratio by filling the target size and cropping excess, 'fit' preserves aspect ratio by fitting within target bounds (may be smaller), 'stretch' forces exact dimensions (may distort). Use 'crop' for thumbnails/avatars, 'fit' for constrained display, 'stretch' for exact sizing.

Crop an image to a specific rectangular region. Specify coordinates and dimensions to extract. Example: crop_image(input='photo.png', x=0, y=100, width=800, height=600, output='cropped.png') extracts an 800x600 region starting at position (0,100). IMPORTANT: All parameters (x, y, width, height) are positional integers, not flags. Coordinates start at (0,0) in the top-left corner. Useful for creating hero images, removing borders, or focusing on specific areas. TIP: Use get_image_info first to check actual image dimensions before cropping.

Compress an image to reduce file size while maintaining visual quality. Quality ranges from 1 (lowest quality, smallest file) to 100 (highest quality, largest file). Default is 90 which provides excellent quality with good compression. Most effective on JPEG images. PNG images are compressed losslessly.

List all available AI image generation providers with pricing, capabilities, and authentication status. Use this FIRST to discover: which providers are configured and ready to use, current per-image pricing, native image-size support, and which credentials are missing. Returns provider IDs (e.g., 'gemini/pro-3', 'vertex/flash-3.1'), pricing, capability flags, and availability status. RECOMMENDED: Call this before generate_image to choose the best provider for your needs.

Scale an image by a factor while preserving aspect ratio. Use this when you want to make an image larger or smaller proportionally. For example, factor 0.5 makes image half size, factor 2.0 makes it double size. Uses high-quality Lanczos resampling.

Resize multiple images in a directory concurrently. Processes all image files (PNG, JPG, WebP, GIF, TIFF, BMP) in the input directory and saves resized versions to the output directory. Uses parallel workers for fast processing of large batches. Supports three modes: 'crop' (default) preserves aspect ratio by filling and cropping, 'fit' preserves aspect ratio by fitting within bounds, 'stretch' forces exact dimensions.

Compress multiple images in a directory concurrently to reduce file sizes. Processes all image files with specified quality setting. Reports total space saved across all images. Uses parallel workers for efficient processing.

Convert multiple images in a directory to a different format concurrently. Useful for converting entire directories to WebP for web optimization, or to PNG for lossless archival. Maintains original filenames with new extensions.

Convert an image between different formats (PNG, JPG/JPEG, WebP, GIF, TIFF, BMP). Useful for web optimization (converting to WebP), compatibility (PNG to JPG), or specific application requirements. Format detection is automatic based on file extension.