google nano banana
Duni
AI Image

Image Editing with Google Nano Banana

Google
August 17, 2026
Wiki 185
promptus ai video generator

Google’s latest AI image model, Gemini Flash 2.5 (aka “Nano Banana”) editing features

Quick answer: Promptus gives you Nano Banana's AI photo editing — local and global edits, multi-turn conversational editing, and photo restoration — directly on your PC through the same desktop app used for local AI image generation, alongside models like Flux and Stable Diffusion, instead of a separate cloud-only editor.

Note: Since this guide was written, Nano Banana has expanded into a model family (Nano Banana 2, Nano Banana Pro, Nano Banana 2 Lite) — see Google's official announcement. Google describes Nano Banana 2 as able to "maintain character resemblance of up to five characters and the fidelity of up to 14 objects in a single workflow," supporting resolutions from 512px to 4K. The core editing workflow below still applies.

Whether you’re restoring an old photo, adjusting lighting, or creating complex multi-turn edits, Nano Banana combines speed, fidelity, and world knowledge to deliver results that feel both natural and precise.

✨ Core Editing Features

1. Local & Global Edits

  • Make changes to a specific part of an image or transform the entire scene.
  • Works seamlessly across small tweaks and large-scale transformations.

Example Prompts:

  • “Remove the person standing in the background.”
  • “Change this photo of the park from daytime to sunset.”

2. Multi-Turn Conversational Editing

gemini flash 2.5
  • Unlike traditional “one-shot” systems, Nano Banana allows step-by-step refinements.
  • Each step builds naturally on the previous edit without breaking image consistency.

Example Workflow:

  • Prompt 1: “Make this a sunset.”
  • Prompt 2: “Add lanterns along the path.”
  • Prompt 3: “Put fireworks in the background.”

3. Recoloring & Restoration

  • Photo restoration: Repair scratches, sharpen blurry details, and colorize black-and-white images.
  • Creative recoloring: Change clothing colors, product finishes, or entire palettes with realistic lighting and textures.

Example Prompts:

  • “Restore this old photo: remove scratches and sharpen the details.”
  • “Recolor this black-and-white wedding photo into natural colors.”
  • “Change the sofa color from beige to emerald green.”

4. Object Addition & Removal

  • Insert new elements into a photo with context-aware realism.
  • Cleanly erase unwanted objects while preserving background detail.

Example Prompts:

  • “Add a coffee cup on the desk next to the laptop.”
  • “Remove the street sign on the left side of the photo.”

5. Pose & Layout Adjustments

  • Modify human or object poses while keeping proportions natural.
  • Works with structured compositions like product catalogs, ID cards, or fashion lookbooks.

Example Prompts:

  • “Adjust the person’s pose so they are sitting cross-legged instead of standing.”
  • “Rearrange the objects on the table into a neat grid layout.”

6. Lighting & Atmosphere Control

  • Change weather, time of day, or mood with natural results.

Example Prompts:

  • “Turn this summer scene into a snowy evening with warm lights inside the house.”
  • “Make this photo look like a cinematic scene with dramatic shadows.”

🎨 Identity & Style Stability

Nano Banana preserves faces, pets, and brand visuals across edits.

Example Prompts:

  • “Keep the same person’s face consistent while changing their outfit into a medieval knight’s armor.”
  • “Apply a watercolor style filter but preserve the character’s identity.”

🛠️ Practical Use Cases

E-Commerce:

  • “Generate a winter holiday version of this shoe catalog photo.”

Architecture:

  • “Replace the brick exterior of this house with modern glass panels.”

Fashion:

  • “Recolor this dress into silk red fabric.”

Heritage Projects:

  • “Restore this faded family portrait and colorize it naturally.”

Marketing:

  • “Create three ad banner variations: tropical theme, winter theme, and minimalist black-and-white.”
google nano banana

🔒 Responsible Editing

  • All edited images include SynthID watermarking.
  • Guardrails prevent harmful or deceptive manipulations.

Example Prompt:

  • “Add branding text to this photo with a visible watermark in the corner.”

📂 Example: Step-by-Step Edit

Original Prompt:

  • “Take this portrait photo. Change background to a sunset beach.”

Follow-up Edits:

  1. “Add string lights behind the subject.”
  2. “Recolor the dress to deep blue velvet.”
  3. “Soften the lighting for a cinematic feel.”

Result:
A natural-looking sequence of edits where the subject’s identity is preserved, and each request builds logically on the last.

👉 With Nano Banana, image editing feels like a conversation — intuitive, precise, and endlessly creative.

Nano Banana
Editing FAQs

It's expanded into a model family — Nano Banana 2, Nano Banana Pro, and Nano Banana 2 Lite — per Google's official announcement. Nano Banana 2 can maintain character resemblance across up to five characters and object fidelity across up to 14 objects in a single workflow, supporting resolutions from 512px to 4K.
Yes — Nano Banana supports multi-turn conversational editing, where each step builds naturally on the previous edit (e.g. \u201cmake this a sunset\u201d, then \u201cadd lanterns\u201d, then \u201cput fireworks in the background\u201d) without breaking image consistency.
Yes — all edited images include SynthID watermarking, and guardrails are in place to prevent harmful or deceptive manipulations.

Nano Banana
Editing FAQs

It's expanded into a model family — Nano Banana 2, Nano Banana Pro, and Nano Banana 2 Lite — per Google's official announcement. Nano Banana 2 can maintain character resemblance across up to five characters and object fidelity across up to 14 objects in a single workflow, supporting resolutions from 512px to 4K.
Yes — Nano Banana supports multi-turn conversational editing, where each step builds naturally on the previous edit (e.g. \u201cmake this a sunset\u201d, then \u201cadd lanterns\u201d, then \u201cput fireworks in the background\u201d) without breaking image consistency.
Yes — all edited images include SynthID watermarking, and guardrails are in place to prevent harmful or deceptive manipulations.
Written by:
Duni
Duni is an Artificial Intelligence engineer at Promptus, specializing in AI workflow design. Duni builds and documents ComfyUI workflows that empower creators to push the boundaries of what’s possible with Promptus and ComfyUI.
Try Promptus Cosy UI today for free.
Start running your first workflow
Go from idea to production-ready output in minutes.
Try Promptus for free ➜