> ## Documentation Index
> Fetch the complete documentation index at: https://docs.znapai.com/llms.txt
> Use this file to discover all available pages before exploring further.
# OpenAI Image generation
***
## Request
```shellscript cURL theme={null}
curl --location 'https://api.znapai.com/images/generations' \
--header 'Content-Type: application/json' \
--header 'Authorization: Bearer $ZnapAI_API_KEY' \
--header 'spend-logs-metadata: {"user_id": "user-123", "project_id": "proj_abc", "env": "production"}' \
--data '{
"model": "gpt-image-2",
"prompt": "Japanese negative film aesthetic, rooftop summer scene, soft natural sunlight, slight overexposure highlights, low contrast, muted faded colors, subtle grain subject standing or sitting on rooftop edge area, body relaxed, slight wind moving hair and clothes, looking toward camera with calm distant gaze, not posing open sky, empty space, minimal elements, imperfect composition, quiet isolated mood, nostalgic and reflective, “memory-like realism” --2:3",
"n": 1,
"size": "1024x1024",
"quality": "low"
}'
```
```python OpenAI SDK (Python) theme={null}
from openai import OpenAI
client = OpenAI(
api_key = "$ZnapAI_API_KEY",
base_url = "https://api.znapai.com/"
)
img = client.images.generate(
model="gpt-image-2",
prompt="Japanese negative film aesthetic, rooftop summer scene, soft natural sunlight, slight overexposure highlights, low contrast, muted faded colors, subtle grain subject standing or sitting on rooftop edge area, body relaxed, slight wind moving hair and clothes, looking toward camera with calm distant gaze, not posing open sky, empty space, minimal elements, imperfect composition, quiet isolated mood, nostalgic and reflective, “memory-like realism” --2:3",
n=1,
size="1024x1024",
quality="low"
)
with open("output.png", "wb") as f:
f.write(image_bytes)
```
```typescript OpenAI SDK (JS) theme={null}
import OpenAI from "openai";
import { writeFile } from "fs/promises";
const client = new OpenAI({
apiKey: "$ZnapAI_API_KEY",
baseURL: "https://api.znapai.com/"
});
const img = await client.images.generate({
model: "gpt-image-2",
prompt: "Japanese negative film aesthetic, rooftop summer scene, soft natural sunlight, slight overexposure highlights, low contrast, muted faded colors, subtle grain subject standing or sitting on rooftop edge area, body relaxed, slight wind moving hair and clothes, looking toward camera with calm distant gaze, not posing open sky, empty space, minimal elements, imperfect composition, quiet isolated mood, nostalgic and reflective, “memory-like realism” --2:3",
n: 1,
size: "1024x1024",
quality: "low"
});
const imageBuffer = Buffer.from(img.data[0].b64_json, "base64");
await writeFile("output.png", imageBuffer);
```
## Response
```json theme={null}
{
"created": 1771066765,
"background": null,
"data": [
{
"b64_json": "iVBORw0KGgoAAAA....",
"revised_prompt": null,
"url": null
}
],
"output_format": "png",
"quality": "low",
"size": "1024x1024",
"usage": {
"total_tokens": 4576,
"input_tokens": 10,
"input_tokens_details": {
"image_tokens": 0,
"text_tokens": 10
},
"output_tokens": 4566,
"output_tokens_details": {
"image_tokens": 4160,
"text_tokens": 406
}
}
}
```
## Image from base64 data
***
# Image generation with additional parameters for supported gemini models
## Request
```shellscript cURL theme={null}
curl --location 'https://api.znapai.com/images/generations' \
--header 'Content-Type: application/json' \
--header 'Authorization: Bearer $ZnapAI_API_KEY' \
--header 'spend-logs-metadata: {"user_id": "user-123", "project_id": "proj_abc", "env": "production"}' \
--data '{
"model": "gemini-3-pro-image",
"prompt": "A panda eating bamboo",
"imageConfig": {
"aspectRatio": "16:9",
"imageSize": "1K"
}
}'
```
## Response
***
## Parameters
An optional header used for spend logging metadata, containing JSON properties like `user_id`, `project_id`, and `env`.
Name of the model to use (for example: gpt-image-2). To see the complete list of supported image generation models, visit: [https://znapai.com/models/image\_generation](https://znapai.com/models/image_generation)
A text description of the desired image(s). The maximum length is 32,000 characters for the GPT image models, 1,000 characters for `dall-e-2` and 4,000 characters for `dall-e-3`.
The number of images to generate. Must be between 1 and 10.
Output image dimensions. Supports arbitrary `WIDTHxHEIGHT` where both edges are multiples of 16, aspect ratio between 1:3 and 3:1, max edge 3840.
Standard values: `1024x1024`, `1536x1024`, `1024x1536`, `auto`.
Supported values depend on the selected model. See the [OpenAI image generation guide](https://developers.openai.com/api/docs/guides/image-generation) for the latest model-specific ranges.
Image quality level. Higher quality increases cost and latency.
Allowed values: `low`, `medium`, `high`, `auto`
Background behavior for the generated image.
Allowed values: `opaque`, `auto`
`transparent` is not supported on gpt-image-2.
Compression level (0–100). Only applies when `output_format` is `jpeg`.
Stream partial images as they are generated. Use with `--no-buffer` in curl.
Number of partial images to emit during streaming. Value between `0` and `3`. Only used when `stream` is `true`.
Content moderation strictness.
Allowed values: `auto`, `low`
Optional configuration parameters for supported Gemini image models.
The aspect ratio of the generated image (e.g. `16:9`, `1:1`, `4:3`).
The resolution level of the image (e.g. `1K`, `2K`, `4K`).
***
## Params to Avoid
| Param | Reason |
| --------------------------- | ------------------------------------------------------------------------------ |
| `response_format: "url"` | ZnapAI image endpoint only returns `b64_json` |
| `background: "transparent"` | Not supported by OpenAI/Gemini image generation models |
| `output_format: "webp"` | WebP output format is not supported |
| `style` | Vivid or natural style configurations are DALL-E 3 exclusive and not supported |
| `spend-logs-metadata` | This metadata must be passed as an HTTP header, not in the JSON request body |