Gemini Nano Banana 2.1 (Stream)

Gemini Nano Banana 2.1 can return generated content through the native Gemini streamGenerateContent endpoint. The endpoint delivers Server-Sent Events (SSE), allowing your application to process response chunks as they arrive.

Key capabilities

SSE streaming — Receive GenerateContentResponse chunks as they are generated.
Text-to-image — Generate images from text prompts.
Image editing — Include a reference image and an editing instruction in the same request.
Image settings — Set output aspect ratio and resolution with imageConfig.

SSE response format

The endpoint returns SSE events. Each data: line contains a JSON GenerateContentResponse chunk. Read candidates[].content.parts[] from each chunk; a part can contain text or image data. Image data uses inlineData with a MIME type and Base64-encoded data. A chunk may also contain usage metadata. The number and ordering of chunks can vary by response.

data: {"candidates":[{"content":{"role":"model","parts":[{"text":"..."}]}}]}

data: {"candidates":[{"content":{"role":"model","parts":[{"inlineData":{"mimeType":"image/png","data":"<base64>"}}]}}]}

data: {"usageMetadata":{"promptTokenCount":24,"candidatesTokenCount":1024,"totalTokenCount":1048}}

Request fields use the native Gemini API format. In a streaming response, generated image bytes are returned in inlineData.data using Base64 encoding.

Text-to-image example

curl --no-buffer \
  "https://www.anyfast.ai/v1beta/models/gemini-nano-banana-2.1:streamGenerateContent?key=YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [{
      "role": "user",
      "parts": [{
        "text": "Create a minimal isometric illustration of a green data center."
      }]
    }],
    "generationConfig": {
      "responseModalities": ["TEXT", "IMAGE"],
      "imageConfig": {"aspectRatio": "16:9", "imageSize": "1K"}
    }
  }'

import base64
import json
import requests

url = "https://www.anyfast.ai/v1beta/models/gemini-nano-banana-2.1:streamGenerateContent"
response = requests.post(
    url,
    params={"key": "YOUR_API_KEY"},
    json={
        "contents": [{
            "role": "user",
            "parts": [{"text": "Create a minimal isometric illustration of a green data center."}],
        }],
        "generationConfig": {
            "responseModalities": ["TEXT", "IMAGE"],
            "imageConfig": {"aspectRatio": "16:9", "imageSize": "1K"},
        },
    },
    stream=True,
)
response.raise_for_status()

for line in response.iter_lines():
    if not line or not line.startswith(b"data:"):
        continue
    chunk = json.loads(line[len(b"data:"):].strip())
    for candidate in chunk.get("candidates", []):
        for part in candidate.get("content", {}).get("parts", []):
            if part.get("thought"):
                continue
            if "text" in part:
                print(part["text"], end="")
            if "inlineData" in part:
                image = base64.b64decode(part["inlineData"]["data"])
                with open("output.png", "wb") as image_file:
                    image_file.write(image)
                print("\nImage saved to output.png")

Image editing example

Send the editing instruction and the reference image in the same parts array. The request uses inlineData with Base64 image bytes.

cURL

curl --no-buffer \
  "https://www.anyfast.ai/v1beta/models/gemini-nano-banana-2.1:streamGenerateContent?key=YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "contents": [{
      "role": "user",
      "parts": [
        {"text": "Keep the product unchanged, replace the background with a warm wooden desk."},
        {"inlineData": {"mimeType": "image/png", "data": "YOUR_BASE64_IMAGE_DATA"}}
      ]
    }],
    "generationConfig": {
      "responseModalities": ["TEXT", "IMAGE"],
      "imageConfig": {"aspectRatio": "4:3", "imageSize": "2K"}
    }
  }'

Parameters

Parameter	Type	Required	Description
`key`	string	Yes	AnyFast API key, passed as a query parameter.
`alt`	string	No	Set to `sse` to explicitly request SSE mode.
`contents`	array	Yes	Conversation turns with text and optional reference media.
`generationConfig.responseModalities`	array	No	Select `IMAGE` or `TEXT` and `IMAGE`.
`generationConfig.imageConfig.aspectRatio`	string	No	Output aspect ratio, such as `1:1`, `4:3`, or `16:9`.
`generationConfig.imageConfig.imageSize`	string	No	Output size: `1K`, `2K`, or `4K`.

API Reference

View the streaming request and response schema for Gemini Nano Banana 2.1.

​Key capabilities

​SSE response format

​Text-to-image example

​Image editing example

​Parameters

API Reference

​Official references

Key capabilities

SSE response format

Text-to-image example

Image editing example

Parameters

Official references