@tasopen/mcp-alphabanana
MCP server for image generation using Google Gemini AI
Versions
1.5.0latest1.4.31.4.21.4.11.4.0+ show 6 moreshow less
1.3.61.3.51.3.41.3.31.3.11.2.1Tools 1
generate_image Generate image assets using Gemini AI with optional transparency and reference images.\n\n[Claude Desktop Guidance]\n- Prefer outputType='file' for medium or large images. base64 and combine responses can exceed Claude Desktop's context limit.\n- On Claude Desktop for Windows, use the FileSystem extension to choose reference-image paths and a writable absolute outputPath before calling this tool.\n- Use base64 only for small previews or when the client explicitly needs inline image data.\n\n[Model Guidance]\n- Flash3.1 (recommended): High quality, very fast, supports grounding and advanced features.\n- Lite3.1 (Nano Banana 2 Lite): Ultra-fast, cost-effective, 1K-only, no search grounding. Ideal for quick drafting and low-latency iteration.\n- Pro3: Higher fidelity, but more costly and slower.\n- Flash2.5: Legacy, maintained for compatibility. Does not support 0.5K, 2K, or 4K resolutions.\n\n[Aspect Ratios]\nGemini supports the following aspect ratios (model-dependent):\n- Common to all models: 1:1, 2:3, 3:2, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 21:9\n- Flash3.1 only: 1:4, 4:1, 1:8, 8:1\n\nNormal mode: provide outputWidth/outputHeight and the server will choose the closest Gemini aspect ratio and source resolution, then resize to the requested pixel size.\nNo-resize mode: set noresize=true and provide aspectRatio plus output_resolution. The server will return Gemini's native pixel dimensions for that combination without post-generation resizing.\n\nIf you intentionally want to control resizing/cropping in normal mode, use the 'resizeMode' parameter: 'crop' (default, center crop), 'letterbox' (fit with padding), 'contain' (trim transparent margins then fit), or 'stretch' (distort to fit).\n\n[IMPORTANT]\nAlways preserve the user's prompt as-is, including language and nuance. Do not translate or summarize.
Permissions 1
env_vars low