First Frame *

Upload First Frame

Last Frame *

Upload Last Frame

Prompt: 0 / 2048

Generate

Sample Video

Nano Banana 2.1 Image to Image: Edit and Restyle Photos

Last verified: November 1, 2026

In Nano Banana 2.1, tiling artifacts are fixed at 2K and 4K on the extreme panoramic ratios: 1:4, 4:1, 1:8 and 8:1. That makes it a good choice when you want to carry an existing photo into a wide banner or a tall vertical strip. Add the source in Upload Images and describe the result you want in the Prompt box, which has room for a full creative brief. Then pick the wide or tall frame under Aspect Ratio and raise Resolution Quality above its default into the range that fix covers. Every other Nano Banana 2.1 Image to Image job works the same way: your images supply the material, and the prompt decides what changes.

Nano Banana 2.1 reads text and images together. Google's model card tests its editing on general edits, stylization, character edits, object and environment edits, doodle-marked edits and multi-image work such as placing several products in a new scene. In practice, one image-to-image form handles both a quick cleanup and a campaign composite. A good working rhythm is to change one thing per run, compare the result with the source, and go again with a sharper instruction.

Explore more image to image

Image To Image About Nano Banana 2.1 Seedream 4.5 Edit Seededit V3 Nano Banana 2 GPT Image 2 Qwen Image 3.0 Pro Qwen Image 3.0

Capability Snapshot

Nano Banana 2.1 Editing Specs at a Glance

Inputs, output options and typical wait time for photo edits on Vidofy.

Input

Up to 10 images (JPG, JPEG, PNG or WEBP, up to 10 MB each) + a text prompt of up to 20,000 characters

Resolution

1K, 2K or 4K

Aspect Ratio

Auto, 1:1, 2:3, 3:2, 1:4, 4:1, 3:4, 4:3, 4:5, 5:4, 1:8, 8:1, 9:16, 16:9, 21:9

Output

1 image per run

Typical Time

About 60 seconds

Prompting Techniques for Better Photo Edits

Describe the Scene You Want, Not What to Leave Out

Google's Nano Banana prompting guide recommends semantic negative prompts: describe the scene you want in positive terms instead of listing what should be absent. In an edit, swap 'no clutter on the desk' for 'a clear oak desk holding only the laptop and a small plant'. That gives the model a concrete target to render instead of a gap to fill. The Prompt box has room for the extra description, so spend it on the outcome rather than on exclusions.

Raise the Resolution for Edits With Small Text

Better text rendering is one of the documented improvements in Nano Banana 2.1. Even so, its model card lists small text as a known weak spot, often blurry in 1K output, along with long paragraphs. When an edit adds a price tag, label or headline, keep the copy short and set Resolution Quality to one of its higher settings for the version you plan to keep. Check the lettering at full size before you download.

How to Edit an Image With Nano Banana 2.1

Five steps, from uploading your source to downloading the edited image.

Add your images in "Upload Images"

Start with the photo you want to change, then add any shots whose people, products or style should appear in the result. "Upload Images" needs an image before the form can run; the FAQ lists the accepted formats and limits. In your prompt, refer to each upload by what it shows, such as 'the woman in the red coat' or 'the green bottle', so the instruction can't be read two ways.

Write the change in the "Prompt" box

The "Prompt" box takes up to 20,000 characters, enough for a complete brief: what changes, what must stay, the style and the light. Start with the change in one plain sentence, then add detail. If you only have a rough idea, the "Enhance Prompt" button under the box suggests fuller versions. Your text only changes when you pick one, so you can compare before committing.

Choose "Aspect Ratio" and "Resolution Quality"

"Aspect Ratio" offers Auto (the default), 1:1, 2:3, 3:2, 1:4, 4:1, 3:4, 4:3, 4:5, 5:4, 1:8, 8:1, 9:16, 16:9 and 21:9. Pick a specific ratio whenever the destination has a fixed shape, such as 9:16 for a story or 4:1 for a site banner. "Resolution Quality" offers 1K (the default), 2K and 4K. Stay at 1K while you're still adjusting the wording, since that is enough to tell whether the edit worked, then move up for the version you keep. Pair the four extreme ratios with 2K or 4K, the resolutions covered by Google's tiling fix.

Click "Generate"

Click "Generate" once the images, prompt and settings are in place. Between runs, change one thing at a time (the wording, the ratio or the resolution) so you can tell which change made the difference.

Download the result

Each run returns one image. Before downloading, check the parts you didn't ask to change, such as edges, hands, faces and any lettering, as closely as the edit itself. If something has drifted, tighten the prompt and generate again rather than settling for a near miss.

Edit Your Own Photos or Start From Text

When the image-editing form fits the job, and when the text-only form of the same model is the better start.

Criterion	Nano Banana 2.1	Alternatives	Best For
What you start from	This form starts from your own images, which the model edits, restyles or combines as the prompt directs.	Text To Image with the same model builds the picture from a written prompt alone, with nothing to upload.	Pick this form when a real photo, product or person has to appear in the result, and Text To Image when the idea so far exists only in words.
Keeping a specific person or product recognizable	Uploaded shots carry the actual look of people and products into the result, within the character and object limits Google documents for multi-image fusion.	Text To Image works from your description, so any person or product has to be conveyed through the words you choose.	Use this form for brand assets and recurring faces that must stay recognizable, and Text To Image for original characters and concept work where you invent the look.

Stay with Nano Banana 2.1 here whenever your own photos need to shape the result, and switch to its Text To Image form when you're inventing the scene from scratch.

Common Nano Banana 2.1 Editing Problems and Fixes

What goes wrong in photo edits, why it happens, and what to change before you run again.

I drew on my photo to mark the area, and the marks are still in the result

The Nano Banana 2.1 model card lists partial instruction following and ink persistence as known limits of masked and doodle-based edits. Hand-drawn marks can survive into the output, and the edit may only half happen. Upload the clean original in Upload Images instead and name the area in the prompt: the object, where it sits in the frame and what it should become. Then run it again.

The person keeps their original pose even though I asked for a new one

The model card notes rare cases where the subject's pose persists during edits, because the structural layout of the input image carries over into the output. Describe the new pose concretely (which way the body faces, where the arms and hands go, the camera angle) and run again. If the whole composition needs to change, generate from a written description in the Text To Image form of the same model, where there is no source structure to hold on to.

My image won't upload

Upload Images checks each file's format, size and pixel dimensions, plus the number of files, against the limits listed in the FAQ. Convert other formats, such as HEIC photos from some phones, to one of the accepted types. Re-export oversized files smaller, replace tiny crops with the full-resolution original, and remove files if you've added too many. Then upload again; there's nothing to rerun.

My full brief doesn't fit in the prompt box

The Prompt box has a character limit, stated in the prompt step above. Cut repetition first, then keep the parts that steer the edit: what changes, what must stay, the style and the light. Drop adjectives that only restate those. Paste the shorter version and generate again.

I only got one image, but I wanted a few options to choose from

Each run on this form produces a single take, so options come from repeat runs. Keep the same uploads and settings, change one part of the wording each time (the style word, the lighting, the background), and compare the takes side by side. Then render your pick again at a higher resolution for the final file.

Frequently Asked Questions

Can I use Nano Banana 2.1 here without a photo to start from?

No. This form only runs once there's an image in Upload Images, because every result here starts from your own material. If your idea exists only in words, the Text To Image form of Nano Banana 2.1 builds the picture from a prompt alone.

Which image files can I upload, and what are the limits?

Upload Images accepts JPG, JPEG, PNG and WEBP files, from 1 to 10 per run. Each file can be up to 10 MB and must be at least 300×300 pixels, so very small crops and thumbnails won't work as sources.

What makes one generation cost more than another?

The credit cost of a run is calculated from the options you choose and shown above the Generate button once you're signed in. Resolution Quality is the setting that changes it, so check the figure after you switch resolution and before you click. Suggestions from the Enhance Prompt button don't use credits.

How long does a Nano Banana 2.1 edit take?

A typical run takes about 60 seconds. That figure is the average of successful generations on Vidofy over the most recent seven days, so any single run can come in a little faster or slower. Use the wait to prepare the next variation of your prompt.