AGRouter

Appearance

Gemini Nano Banana 2.1（Nano Banana 2.1）

POST/v1beta/models/gemini-nano-banana-2.1:generateContent

Generate or edit images with Nano Banana 2.1 through the native Google Gemini protocol. AGRouter preserves the provider request and response payload instead of converting it to Chat Completions.

Request body

Send a native Gemini GenerateContent body. Provider-defined fields are passed through for this model.

arraycontentsrequired

Gemini Content objects containing the prompt and optional input images.

array<string>generationConfig.responseModalitiesrequired

Requested output modalities. Include IMAGE; include TEXT when you also want a text part.

Optional<string>generationConfig.imageConfig.aspectRatio

Output aspect ratio, for example 1:1, 16:9, 9:16, 21:9, 4:1 or 1:8.

Optional<string>generationConfig.imageConfig.imageSize

Output resolution: 1K, 2K or 4K. Nano Banana 2.1 has no 0.5K tier.

Edit an image ​

Add the source image as an inlineData part next to the text instruction. Use the MIME type of the bytes you send:

json
{
  "contents": [{
    "role": "user",
    "parts": [
      {"text": "Keep the subject unchanged and turn the background into a paper-cut landscape."},
      {"inlineData": {"mimeType": "image/png", "data": "<base64-input-image>"}}
    ]
  }],
  "generationConfig": {"responseModalities": ["IMAGE"], "imageConfig": {"imageSize": "1K"}}
}

Use the Google Gen AI SDK ​

Point the official SDK at AGRouter and use your AGRouter API key:

python
from google import genai
from google.genai import types

client = genai.Client(
    api_key="YOUR_AGROUTER_API_KEY",
    http_options=types.HttpOptions(base_url="https://api.agrouter.ai"),
)

response = client.models.generate_content(
    model="gemini-nano-banana-2.1",
    contents="Create a wide editorial illustration of a solar-powered city at sunrise.",
    config=types.GenerateContentConfig(
        response_modalities=["IMAGE"],
        image_config=types.ImageConfig(aspect_ratio="16:9", image_size="1K"),
    ),
)

for part in response.candidates[0].content.parts:
    if part.inline_data:
        with open("output.jpg", "wb") as f:
            f.write(part.inline_data.data)

For streaming, call :streamGenerateContent?alt=sse (SDK: generate_content_stream) with the same body.

Response handling ​

Read generated media from the returned candidate parts. Because this model uses native passthrough, provider-defined field casing and MIME metadata are returned as received. The output is usually JPEG; decode the base64 payload and verify the actual media bytes before choosing a file extension. Keep thoughtSignature when you send the turn back in multi-turn editing.

Official Google resources ​