Nano Banana 2.1 API on MuAPI

Generate and edit images with Google's Nano Banana 2.1 through one REST API. This repository is a developer guide with runnable examples for the Nano Banana 2.1 text-to-image and image-edit endpoints on MuAPI. It is an independent integration guide and is not an official Google project.

• MuAPI: https://muapi.ai
• Nano Banana 2.1 API page (endpoints, comparison, FAQ): https://muapi.ai/nano-banana-2.1
• Text to image playground: https://muapi.ai/playground/nano-banana-2-1
• Image edit playground: https://muapi.ai/playground/nano-banana-2-1-edit
• All Nano Banana endpoints: https://muapi.ai/nano-banana-api
• Create/manage an API key: https://muapi.ai/access-keys
• API introduction: https://muapi.ai/docs/introduction
• Pricing: https://muapi.ai/pricing

Related Projects

• MuAPI — Unified API for image, video, and audio generation across hundreds of AI models.
• Nano Banana 2.1 on MuAPI — Landing page, endpoint comparison, and FAQ for this repository's API.
• Open Generative AI — Open-source generative-media app and model references.
• Nano Banana Generator — Open-source Next.js app for Nano Banana text-to-image and editing.
• Awesome AI Image Models — Image model and API comparison hub.
• Google Gemini Media API — Image and video generation API examples across Google models.
• Qwen Image API — Another image generation and editing API guide on the same MuAPI task API.

What Nano Banana 2.1 is

Nano Banana 2.1 is Google's image generation and editing model on the Flash tier, the successor to Nano Banana 2 and Nano Banana Pro. It accepts text and image inputs and returns images, with 1K, 2K, and 4K output and a wide range of aspect ratios. Google describes improvements over its earlier models in visual design, mask-based editing, and subject consistency (announcement).

On MuAPI it is exposed as two endpoints that share the same authentication, task lifecycle, and result polling as every other model.

Video tutorial

Watch the two-minute walkthrough: How to Use the Nano Banana 2.1 API (Step-by-Step Guide). It covers what Google says is new, creating an API key, sending a request, checking back with the request ID, and editing your own image. Creator tests shown in the video belong to their authors and were not generated by MuAPI.

Choose an endpoint

Workflow	Endpoint	What it does	Required input
Text to image	nano-banana-2-1	Generate an image from a prompt	prompt
Image edit	nano-banana-2-1-edit	Edit or combine reference images with a plain-language instruction	prompt, images_list

Want to compare against the earlier models first? They use the same request shape:

Model	Endpoint	Tier	Best for
Nano Banana 2.1	nano-banana-2-1 / nano-banana-2-1-edit	Flash	Latest Flash-tier generation and editing
Nano Banana 2	nano-banana-2 / nano-banana-2-edit	Flash	Established default
Nano Banana Pro	nano-banana-pro / nano-banana-pro-edit	Pro	Highest-fidelity professional work

Check muapi.ai/pricing and the model playground for current pricing and availability before production use. Prices are intentionally not copied into this README so it cannot go stale.

How the API works

1. Submit POST https://api.muapi.ai/api/v1/{endpoint} with an x-api-key header and a JSON body.
2. Read the returned request_id (it is like a receipt for your image).
3. Poll GET https://api.muapi.ai/api/v1/predictions/{request_id}/result until the status is completed, then read the image URL from the result.

Quick start: generate an image

export MUAPI_KEY="your_muapi_api_key"

curl -X POST "https://api.muapi.ai/api/v1/nano-banana-2-1" \
  -H "x-api-key: $MUAPI_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "A ceramic tea set on a sunlit kitchen table, soft morning light",
    "aspect_ratio": "1:1",
    "resolution": "2k",
    "output_format": "png"
  }'

Then fetch the result with the returned request_id:

curl "https://api.muapi.ai/api/v1/predictions/REQUEST_ID/result" \
  -H "x-api-key: $MUAPI_KEY"

Image edit example

images_list is a list of publicly reachable image URLs (the API receives URLs; it does not upload local files). Image order matters when the prompt refers to the first or second image.

curl -X POST "https://api.muapi.ai/api/v1/nano-banana-2-1-edit" \
  -H "x-api-key: $MUAPI_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "Change the jacket to a teal windbreaker and keep everything else the same",
    "images_list": ["https://example.com/photo.jpg"],
    "resolution": "2k"
  }'

Python example

Requires Python 3.9+ and requests. The script submits a generate or edit request and polls until the image is ready.

python -m pip install -r requirements.txt
export MUAPI_KEY="your_muapi_api_key"

python examples/nano_banana_21.py --prompt "A tiny greenhouse on a rainy city rooftop at dusk"

python examples/nano_banana_21.py --mode edit \
  --prompt "Turn the product into a brushed aluminum desk lamp; keep the composition" \
  --image-url "https://example.com/product.jpg"

JavaScript example

Requires Node 18+ (built-in fetch). See examples/nano_banana_21.mjs.

export MUAPI_KEY="your_muapi_api_key"
node examples/nano_banana_21.mjs "A tiny greenhouse on a rainy city rooftop at dusk"

Request fields

These follow the Nano Banana family contract on MuAPI. Confirm the current form for 2.1 in the playground before relying on a field.

Field	Type	Required	Accepted values / behavior
prompt	string	yes	What to generate, or the edit instruction
images_list	array of URLs	edit only	Reference images for the edit endpoint
aspect_ratio	string	no	1:1, 1:4, 1:8, 2:3, 3:2, 3:4, 4:1, 4:3, 4:5, 5:4, 8:1, 9:16, 16:9, 21:9, or Auto
resolution	string	no	1k, 2k, or 4k (lowercase)
output_format	string	no	jpg or png
google_search	boolean	no	Use Google Search to enhance the prompt

Scope and honesty notes

• This repository documents how to call the endpoints through MuAPI. It does not host or modify the model.
• Model capabilities described here come from Google's public announcement; MuAPI makes no benchmark claims in this guide. Compare outputs on your own prompts in the playground.
• Endpoint availability and inputs can change; the MuAPI Nano Banana 2.1 page is the source of truth.

License

MIT, see LICENSE.