Nano Banana 2 Lite Generation Workflow

Plan Nano Banana 2 Lite image generations with prompts, source images, settings, credits, small iterations and downloads.

Previous: Reference Image Editing · Next: Troubleshooting

Choose Input Mode

Three-panel workflow illustration showing an upload area, an AI generation canvas and an export preview.
  1. 1

    Choose input mode

    Text-to-image starts from a written prompt alone; image-to-image starts from an uploaded source image plus a prompt.

  2. 2

    Define your image brief

    Write the subject, composition, lighting and style in one or two sentences — see the Prompt Guide for the full formula.

  3. 3

    Upload clean source files

    For image-to-image, upload a sharp, well-lit source image with one clear subject — see Reference Image Editing for the full quality checklist.

  4. 4

    Set aspect ratio & Native 1K output

    Pick an aspect ratio for your composition. Every option generates at Native 1K — see the output table below for what that means per ratio.

  5. 5

    Review credits before you submit

    Check the credit cost shown before submitting — text-to-image and image-to-image are priced separately. See Pricing for current credit costs.

  6. 6

    Iterate in small changes

    Change one detail of your prompt at a time between attempts rather than rewriting it completely — it is the fastest way to converge on the result you want.

  7. 7

    Save and download

    Once a result matches your brief, download it. Nothing is auto-deleted from your session, but downloading keeps a permanent copy.

Native 1K Output by Aspect Ratio

This is the only page on the site that lists per-aspect-ratio output size — every other page links back here instead of repeating numbers.

Native 1K output by aspect ratio. Exact size is confirmed only for 1:1; all other ratios keep roughly the same pixel budget on their long edge — treat your actual generation result as authoritative.
Aspect ratioApprox. output sizeNotes
1:11024 × 1024 pxConfirmed native baseline — square
3:2long edge ≈ 1024pxLandscape — exact dimensions vary by generation, treat the actual output as authoritative
2:3long edge ≈ 1024pxPortrait — exact dimensions vary by generation, treat the actual output as authoritative
16:9long edge ≈ 1024pxWidescreen landscape — exact dimensions vary by generation, treat the actual output as authoritative
9:16long edge ≈ 1024pxWidescreen portrait — exact dimensions vary by generation, treat the actual output as authoritative

Data last reviewed 2026-07-05 against Google's official Gemini image generation docs and this product's own Replicate integration notes.

Ready to put this into practice? Start at Text to Image for a prompt-only generation, or Image to Image if you have a source photo to edit — check current credit costs on Pricing before you submit.

FAQ

Start Creating

Put the workflow into practice with your own prompt.