Text-to-Image

Generate an image from a text prompt with Nano Banana 2 Lite. The model ID is gemini-3.1-flash-lite-image, without a -preview suffix.

This model supports 1K output only and has an independent price and API-key permission. Enabling Nano Banana 2 does not automatically enable Nano Banana 2 Lite. Do not request 2K or 4K; those sizes are not configured for this model.

POST /api/v1/generations

This is the asynchronous native API: a successful request returns 202 Accepted. Poll GET /api/v1/jobs/{job_id} to retrieve the result. For a synchronous Gemini-compatible call, see the SDK quickstart.

Headers

Header	Required	Description
`Authorization`	Yes	`Bearer <MOONEZ_API_KEY>`. The key must have permission for Nano Banana 2 Lite.
`Content-Type`	Yes	`application/json`
`Idempotency-Key`	No	Repeating the same key returns the original `job_id`; see Authentication.

Body

Field	Required	Description
`model`	Yes	`gemini-3.1-flash-lite-image`
`prompt`	Yes	The prompt, at least one character long.
`aspect_ratio`	Yes	Aspect ratio, such as `16:9` or `auto`.
`resolution`	Yes	Resolution: `1K` only.
`callback_url`	No	Webhook URL for this job.
`metadata`	No	An arbitrary JSON object.
`timeout`	No	Total execution budget in seconds, from 30 to 1200.

With aspect_ratio: "auto", Moonez.ai does not specify an aspect ratio to the provider. This value applies to the native API; in the Gemini-compatible API, simply omit generationConfig.imageConfig.aspectRatio.

For text-only input, omit input_images_urls and input_images_base64. For editing, see Image-to-image.

Request examples

Set your key in the MOONEZ_API_KEY environment variable.

curl -sS https://api.moonez.ai/api/v1/generations \
  -H "Authorization: Bearer $MOONEZ_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-3.1-flash-lite-image",
    "prompt": "Draw a red panda astronaut against a nebula.",
    "aspect_ratio": "16:9",
    "resolution": "1K"
  }'

import os

import requests

response = requests.post(
    "https://api.moonez.ai/api/v1/generations",
    headers={"Authorization": f"Bearer {os.environ['MOONEZ_API_KEY']}"},
    json={
        "model": "gemini-3.1-flash-lite-image",
        "prompt": "Draw a red panda astronaut against a nebula.",
        "aspect_ratio": "16:9",
        "resolution": "1K",
    },
    timeout=30,
)
response.raise_for_status()
print(response.json())

const response = await fetch("https://api.moonez.ai/api/v1/generations", {
  method: "POST",
  headers: {
    Authorization: `Bearer ${process.env.MOONEZ_API_KEY}`,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    model: "gemini-3.1-flash-lite-image",
    prompt: "Draw a red panda astronaut against a nebula.",
    aspect_ratio: "16:9",
    resolution: "1K",
  }),
  signal: AbortSignal.timeout(30_000),
});
if (!response.ok) throw new Error(`Generation request failed: ${response.status}`);
console.log(await response.json());

Response

202 Accepted means the job has been accepted, not that generation is complete.

{
  "job_id": "0193a7f2-1c4a-7e0b-9b3a-2f1d8c6e4a55",
  "status": "queued",
  "status_url": "/api/v1/jobs/0193a7f2-1c4a-7e0b-9b3a-2f1d8c6e4a55"
}

Poll status_url with the same Authorization header until the status is done or failed. If the model returns an image, its link is available in result.image_url. The model may also return text; check the result contents.

See API errors for 401, 402, 403, 422, and 503 responses.