Regen API

The Regen API is an API that provides functionality to regenerate masked parts of an image using AI. When an empty string ("") is passed as the prompt, the focus is on removing the objects within the masked area to blend naturally with the surroundings. However, if something is written in the prompt, the focus shifts to adding that object, modifying the image accordingly.

---

Authentication

header

X-Photio-Key : {PHOTIO_API_KEY}

Do you not have an API key yet?

To call the API, you first need to obtain a Photio API key. To get your API key, go to the API Dashboard. After registering your card, you can use the issued API key.

---

Request Body

Variable	Type	Require	Default	Description
prompt	string	✓	""	This is a description of the image you want to create, provided in English. It should include a detailed depiction of the scene or object you'd like to generate.
base_image	string	✓	-	This can be either a base image in base64 format or an image URL. The output generated by the AI will match the dimensions of the input base image (e.g., input size 1024*1024 will produce output of the same size). The recommended pixel size for images is as below
mask_image	string	✓	-	This can be either a mask_image image in base64 format or an image URL. The mask_image has to be same size with base_image. Mask black(0) would be edited with AI.
negative_prompt	string		"worst quality, low quality, illustration, 3d, 2d, painting, cartoons, sketch"	Enter in English any objects or scenarios you do not want to appear in the image. If nothing is entered, a default negative prompt will be used, which includes: "worst quality, low quality, illustration, 3d, 2d, painting, cartoons, sketch."
quality	string		"LOW"	Choose from three levels of quality: "LOW", "MIDDLE", "HIGH". The default is "LOW". Setting the quality to "HIGH" will increase the detail quality of the generated image but will also take more time. A "LOW" quality setting results in relatively lower detail quality but faster generation speed.
seed	int		-1	This variable allows for the creation of different outcomes even with the same input. It can be set to any integer between 1 and 4,294,967,296. If you enter -1 or leave this field blank, a random seed number will be used to generate the image.
cfg_scale	float		2.5	This setting determines how closely the AI should follow the prompt. A higher cfg_scale (e.g., 2.5) means the AI will strictly adhere to the prompt and not add extraneous elements. Lower cfg_scale settings allow the AI more freedom to incorporate various elements into the image. Recommend value range is 1.0~2.5

Shell

curl --request POST
--url https://api.photio.io/v1/inference/regen
--header 'X-Photio-Key: {X-Photio-Key}'
--header 'Content-Type: application/json'
--data '{ "prompt": "starfish", "mask_image": "https://d2yiwnckkg6ukk.cloudfront.net/example/mask.png", "base_image": "https://d2yiwnckkg6ukk.cloudfront.net/example/perfume_on_the_beach_output.png" }'

Python

import requests
import json

data = {
  "prompt": "starfish",
  "mask_image": "https://d2yiwnckkg6ukk.cloudfront.net/example/mask.png",
  "base_image": "https://d2yiwnckkg6ukk.cloudfront.net/example/perfume_on_the_beach_output.png"
}

body_data = json.dumps(data)

# Set the necessary headers
headers = {
  'X-Photio-Key': '{X-Photio-Key}',
  'Content-Type': 'application/json'
}

response = requests.post('https://api.photio.io/v1/inference/regen', headers=headers, data=body_data)

Java

class Generate {
  public void remix() {
    URL url = new URL("https://api.photio.io/v1/inference/regen");
    HttpURLConnection conn = (HttpURLConnection) url.openConnection();

    conn.setRequestProperty("X-Photio-Key", "${PHOTIO_API_KEY}");
    conn.setRequestProperty("Content-Type", "application/json");

    conn.setDoOutput(true);
    conn.setDoInput(true);

    conn.setRequestMethod("POST");

    JSONObject data = new JSONObject();
    data.put("prompt", "starfish");
    data.put("mask_image", "https://d2yiwnckkg6ukk.cloudfront.net/example/mask.png");
    data.put("base_image", "https://d2yiwnckkg6ukk.cloudfront.net/example/perfume_on_the_beach_output.png");

    String bodyData = data.toString();

    try (OutputStream os = conn.getOutputStream()) {
      byte[] input = bodyData.getBytes("utf-8");
      os.write(input, 0, input.length);
    }

    int responseCode = conn.getResponseCode();
    System.out.println("Response Code : " + responseCode);

    conn.disconnect();
  }
}

JS

const data = {
  prompt: "starfish",
  mask_image: "https://d2yiwnckkg6ukk.cloudfront.net/example/mask.png",
  base_image: "https://d2yiwnckkg6ukk.cloudfront.net/example/perfume_on_the_beach_output.png"
};

const bodyData = JSON.stringify(data);

// Set the necessary headers
const headers = {
'X-Photio-Key': '{X-Photio-Key}',
'Content-Type': 'application/json'
};

fetch('https://api.photio.io/v1/inference/regen', {
method: 'POST',
headers: headers,
body: bodyData
})
.then(response => response.json())
.then(data => console.log(data))
.catch(error => console.error('Error:', error));

Recommended image pixel size

Tag	Ratio	Pixel size
Fullscreen	4:3	1152x896
Widescreen	16:9	1344x768
Ultrawide	21:9	1536x640
Mobile landscape	3:2	1216x832
Square	1:1	1024x1024
Mobile Portrait	2:3	832x1216
Tall	9:16	768x1344

Response

Variable	Type	Description
status_code	int	In the response, you can find a status value to check the status of the request. A normal response is indicated by 200, authentication issues by 401, and API status problems by a 500 error.
body	string	The result image data is output as a Base64-encoded string.

Response Example

{
    "status_code": Int,
    "body": Base64 String
}

Output Example