Nano Banana 2.1: Image Edit
Edit images with natural language and mask-based local edits powered by Nano Banana 2.1, taking one input image and returning one edited result alongside a side-by-side comparison view. Ideal for brand marketing posters, e-commerce product shots, and multi-round retouching where subjects and text must stay consistent.
About
This workflow compares two editing strategies in Nano Banana 2.1—prompt-only global edits versus mask-guided local edits—using a single model node (GeminiNanoBanana2V2) and a consistent pipeline. It loads your image (e.g., the provided silver_frame_profile.png), applies your natural-language edit, and renders a side-by-side before/after via ImageCompare so you can judge impact at a glance. Running once without a mask showcases broad, scene-wide transformations; running with a mask constrains changes to selected regions while preserving the rest.
What you provide
- Reading the workflow
Runs in ComfyUI. This workflow can't run on this page. Open it in Comfy Cloud with everything loaded, or download it for your own ComfyUI.
Try in CloudQuestions & answers
The same edit prompt is applied with and without a mask using the GeminiNanoBanana2V2 node (Nano Banana 2.1). The no-mask run shows global changes, while the masked run restricts edits to selected regions. ImageCompare visualizes before/after so you can evaluate control versus cohesion.
Use masks when subject identity, product details, or text must remain stable. Masking limits unintended changes and drift. Choose global edits when you want coherent style or mood shifts across the entire image and minor variations are acceptable.
Keep the prompt and all GeminiNanoBanana2V2 parameters identical across runs and change only the mask input. If your setup exposes a seed, fix it to reduce randomness. Save outputs with clear names and document settings in MarkdownNote.
Hard mask boundaries can create visible transitions. Use a softened/feathered mask edge, ensure color harmony in your prompt, and consider slightly expanding or refining the mask to blend the edited and protected regions more naturally.