> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://developer.ideogram.ai/v1/api-reference/describe-prompts/magic-prompt-v4/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://developer.ideogram.ai/_mcp/server.

# Generate a Magic Prompt with Ideogram 4.0

POST https://api.ideogram.ai/v1/ideogram-v4/magic-prompt
Content-Type: application/json

Transforms a basic prompt into an enhanced Ideogram 4.0 magic prompt.
The magic prompt model version is fixed; callers cannot select it.
When `aspect_ratio` is `AUTO` (the default), the model selects the most
suitable aspect ratio for the prompt and returns it in the response.


Reference: https://developer.ideogram.ai/v1/api-reference/describe-prompts/magic-prompt-v4

## Authentication

- `Api-Key` header (required) — API key for access control. Use in the header with the name \"Api-Key\"

## Request

### Body (application/json)

This endpoint expects a MagicPromptV4Request.

- `text_prompt` (string, required) — The natural-language prompt to enhance into an Ideogram 4.0 magic prompt. Named `text_prompt` to match the Ideogram 4.0 generate request surface.
- `aspect_ratio` (enum, optional, default: AUTO) — Target aspect ratio. Defaults to `AUTO`, which lets the model select the most suitable ratio and return it in the response.
  - Allowed values: `AUTO`, `1x4`, `1x3`, `1x2`, `9x16`, `10x16`, `2x3`, `3x4`, `4x5`, `1x1`, `5x4`, `4x3`, `3x2`, `16x10`, `16x9`, `2x1`, `3x1`, `4x1`

## Response

### 200

Magic prompt generated successfully.

- `json_prompt` (V4JsonPrompt, required) — The enhanced prompt as an Ideogram 4.0 structured JSON prompt. This conforms to the same contract as `json_prompt` on `/v1/ideogram-v4/generate`, so it can be passed straight back to that endpoint.
- `aspect_ratio` (enum, required, default: AUTO) — The resolved aspect ratio. When the request used `AUTO`, this is the ratio the model selected; otherwise it echoes the request.
  - Allowed values: `AUTO`, `1x4`, `1x3`, `1x2`, `9x16`, `10x16`, `2x3`, `3x4`, `4x5`, `1x1`, `5x4`, `4x3`, `3x2`, `16x10`, `16x9`, `2x1`, `3x1`, `4x1`

## Errors

### 400 Bad Request Error

Invalid input provided.

- `any`

### 401 Unauthorized Error

Not authorized to generate magic prompts.

- `any`

### 429 Too Many Requests Error

Too many requests.

- `any`

### 500 Internal Server Error

Internal server error.

- `any`

## Types

### V4JsonPrompt

Structured prompt for Ideogram 4.0 generation. When `json_prompt` is supplied, magic-prompt is disabled and the diffusion model consumes the JSON contract directly. Mutually exclusive with `text_prompt` and the legacy `prompt` field.

- `high_level_description` (string, required) — One- or two-sentence overall description of the desired image.
- `compositional_deconstruction` (V4CompositionalDeconstruction, required) — The compositional breakdown of a V4 prompt — background plus an ordered list of elements.
- `style_description` (V4StyleDescription, optional) — Optional style description supplied alongside a V4 JSON prompt.

### V4CompositionalDeconstruction

The compositional breakdown of a V4 prompt — background plus an ordered list of elements.

- `background` (string, required) — Description of the background of the scene.
- `elements` (list of V4PromptElement, required) — Ordered list of elements (objects and text) composing the scene.

### V4StyleDescription

Optional style description supplied alongside a V4 JSON prompt.

- `aesthetics` (string, optional) — Aesthetic notes (mood, vibe, references).
- `art_style` (string, optional) — Optional art-style hint (e.g., illustration, oil painting).
- `lighting` (string, optional) — Lighting description.
- `medium` (string, optional) — Medium description (e.g., photograph, digital art).
- `photo` (string, optional) — Optional photographic style notes (e.g., lens, film stock).
- `color_palette` (list of string, optional) — Optional list of hex color strings (e.g. "#FFD700") that bias the Ideogram 4.0 output toward this palette. Applied as a soft color bias, not an exact per-pixel lock.

### V4PromptElement

A single element in the V4 prompt's compositional deconstruction. Discriminated by the `type` field.

- `type`: `obj` (V4ObjPromptElement)
  - `desc` (string, required) — Description of the object element.
  - `bbox` (list of integer, optional) — Bounding box for an element, expressed as four integers in `[0, 1000]` with `[y_min, x_min, y_max, x_max]` semantics (row-first). Values are normalized so the canvas is `1000 x 1000` regardless of the final resolution.
  - `color_palette` (list of string, optional) — Optional list of hex color strings (e.g. "#FFD700") that bias the Ideogram 4.0 output toward this palette. Applied as a soft color bias, not an exact per-pixel lock.
- `type`: `text` (V4TextPromptElement)
  - `desc` (string, required) — Description of the text element (style, role, placement notes).
  - `text` (string, required) — The literal text to render in the image.
  - `bbox` (list of integer, optional) — Bounding box for an element, expressed as four integers in `[0, 1000]` with `[y_min, x_min, y_max, x_max]` semantics (row-first). Values are normalized so the canvas is `1000 x 1000` regardless of the final resolution.
  - `color_palette` (list of string, optional) — Optional list of hex color strings (e.g. "#FFD700") that bias the Ideogram 4.0 output toward this palette. Applied as a soft color bias, not an exact per-pixel lock.

## Examples

**Request**

```json
{
  "text_prompt": "A photo of a cat",
  "aspect_ratio": "AUTO"
}
```

**Response**

```json
{
  "json_prompt": {
    "high_level_description": "A photorealistic ginger cat perched on a vintage wooden chair by a sunlit window.",
    "compositional_deconstruction": {
      "background": "A dim room with a bright window casting warm light.",
      "elements": [
        {
          "type": "obj",
          "desc": "ginger cat with green eyes sitting upright on a vintage wooden chair"
        }
      ]
    },
    "style_description": {
      "aesthetics": "warm, cozy, nostalgic",
      "lighting": "soft natural window light",
      "medium": "photograph"
    }
  },
  "aspect_ratio": "1x1"
}
```

**SDK Code**

```python
import requests

url = "https://api.ideogram.ai/v1/ideogram-v4/magic-prompt"

payload = {
    "text_prompt": "A photo of a cat",
    "aspect_ratio": "AUTO"
}
headers = {
    "Api-Key": "<apiKey>",
    "Content-Type": "application/json"
}

response = requests.post(url, json=payload, headers=headers)

print(response.json())
```

```javascript
const url = 'https://api.ideogram.ai/v1/ideogram-v4/magic-prompt';
const options = {
  method: 'POST',
  headers: {'Api-Key': '<apiKey>', 'Content-Type': 'application/json'},
  body: '{"text_prompt":"A photo of a cat","aspect_ratio":"AUTO"}'
};

try {
  const response = await fetch(url, options);
  const data = await response.json();
  console.log(data);
} catch (error) {
  console.error(error);
}
```

```go
package main

import (
	"fmt"
	"strings"
	"net/http"
	"io"
)

func main() {

	url := "https://api.ideogram.ai/v1/ideogram-v4/magic-prompt"

	payload := strings.NewReader("{\n  \"text_prompt\": \"A photo of a cat\",\n  \"aspect_ratio\": \"AUTO\"\n}")

	req, _ := http.NewRequest("POST", url, payload)

	req.Header.Add("Api-Key", "<apiKey>")
	req.Header.Add("Content-Type", "application/json")

	res, _ := http.DefaultClient.Do(req)

	defer res.Body.Close()
	body, _ := io.ReadAll(res.Body)

	fmt.Println(res)
	fmt.Println(string(body))

}
```

```ruby
require 'uri'
require 'net/http'

url = URI("https://api.ideogram.ai/v1/ideogram-v4/magic-prompt")

http = Net::HTTP.new(url.host, url.port)
http.use_ssl = true

request = Net::HTTP::Post.new(url)
request["Api-Key"] = '<apiKey>'
request["Content-Type"] = 'application/json'
request.body = "{\n  \"text_prompt\": \"A photo of a cat\",\n  \"aspect_ratio\": \"AUTO\"\n}"

response = http.request(request)
puts response.read_body
```

```java
import com.mashape.unirest.http.HttpResponse;
import com.mashape.unirest.http.Unirest;

HttpResponse<String> response = Unirest.post("https://api.ideogram.ai/v1/ideogram-v4/magic-prompt")
  .header("Api-Key", "<apiKey>")
  .header("Content-Type", "application/json")
  .body("{\n  \"text_prompt\": \"A photo of a cat\",\n  \"aspect_ratio\": \"AUTO\"\n}")
  .asString();
```

```php
<?php
require_once('vendor/autoload.php');

$client = new \GuzzleHttp\Client();

$response = $client->request('POST', 'https://api.ideogram.ai/v1/ideogram-v4/magic-prompt', [
  'body' => '{
  "text_prompt": "A photo of a cat",
  "aspect_ratio": "AUTO"
}',
  'headers' => [
    'Api-Key' => '<apiKey>',
    'Content-Type' => 'application/json',
  ],
]);

echo $response->getBody();
```

```csharp
using RestSharp;

var client = new RestClient("https://api.ideogram.ai/v1/ideogram-v4/magic-prompt");
var request = new RestRequest(Method.POST);
request.AddHeader("Api-Key", "<apiKey>");
request.AddHeader("Content-Type", "application/json");
request.AddParameter("application/json", "{\n  \"text_prompt\": \"A photo of a cat\",\n  \"aspect_ratio\": \"AUTO\"\n}", ParameterType.RequestBody);
IRestResponse response = client.Execute(request);
```

```swift
import Foundation

let headers = [
  "Api-Key": "<apiKey>",
  "Content-Type": "application/json"
]
let parameters = [
  "text_prompt": "A photo of a cat",
  "aspect_ratio": "AUTO"
] as [String : Any]

let postData = JSONSerialization.data(withJSONObject: parameters, options: [])

let request = NSMutableURLRequest(url: NSURL(string: "https://api.ideogram.ai/v1/ideogram-v4/magic-prompt")! as URL,
                                        cachePolicy: .useProtocolCachePolicy,
                                    timeoutInterval: 10.0)
request.httpMethod = "POST"
request.allHTTPHeaderFields = headers
request.httpBody = postData as Data

let session = URLSession.shared
let dataTask = session.dataTask(with: request as URLRequest, completionHandler: { (data, response, error) -> Void in
  if (error != nil) {
    print(error as Any)
  } else {
    let httpResponse = response as? HTTPURLResponse
    print(httpResponse)
  }
})

dataTask.resume()
```