A VS Code extension that generates images with Gemini API.
It can build prompts for image generation through the GitHub Copilot Language Model API, then generate and open the output image in VS Code.
Features
Right-click command from editor selection: Generate Image with Nano Banana Pro
Optional final refinement step: edit the just-generated image with an extra instruction (reference image based)
Right-click on image (.png, .jpg, .jpeg) in Editor or Explorer: Edit Image with Nano Banana Pro
Open Nano Banana Image Editor from image context menu for visual annotations (rectangle, circle, text) + prompt-based editing
Secure Gemini API key storage (SecretStorage)
Style picker (12 built-in styles)
Aspect ratio picker
Configurable image size (1K, 2K, 4K)
Cancellable generation progress UI
Prompt generation with VS Code Copilot LM API (vscode.lm)
Runtime i18n for extension messages
Supported UI Languages
English (en)
Traditional Chinese (zh-TW)
Simplified Chinese (zh-CN)
Japanese (ja)
Korean (ko)
Thai (th)
Vietnamese (vi)
Language behavior:
Default: auto-detect from current VS Code display language
Optional override: nanoBanana.displayLanguage
Commands
nanoBanana.generateFromSelection
nanoBanana.setGeminiApiKey
nanoBanana.selectCopilotPromptModel (loads currently available models from VS Code Language Model API)
nanoBanana.editImageWithReference (edit selected/open image as a Gemini reference image)
nanoBanana.openImageEditor (open the visual annotation editor for image-to-image editing)
Command labels are localized through package.nls.*.json.