Magic Prompt AI

1 tokenTools

Transform images using text prompts

Overview

Magic Prompt AI lets you transform any image using natural language instructions. Describe what changes you want and the AI applies them — change colors, add elements, modify materials, or completely restyle a scene. You can provide up to 6 reference images to guide the style and composition.

Image input

Endpoint

POST /api/v1/generate

bash

curl -X POST https://roomai.com/api/v1/generate \
  -H "Authorization: Bearer rai_live_your_api_key" \
  -H "Content-Type: application/json" \
  -d '{"model": "magicprompt", ...}'

Parameters

Parameter	Type	Required	Description
`model`	string	Required	Must be "magicprompt". magicprompt
`image`	string	Required	URL of the base image to transform.
`prompt`	string	Required	Describe the transformation (e.g. "change the walls to exposed brick and add warm pendant lighting").
`precise`	string	Optional	AI strength from "0" to "100". Higher values stay closer to the original image. Default: `"50"`
`enhance_prompt`	string	Optional	Set to "true" to let the AI enhance your prompt. Default: `"false"` truefalse
`extra_image_1`	string	Optional	URL of reference image 1 (up to 6 supported).
`extra_image_2`	string	Optional	URL of reference image 2.
`extra_image_3`	string	Optional	URL of reference image 3.
`extra_image_4`	string	Optional	URL of reference image 4.
`extra_image_5`	string	Optional	URL of reference image 5.
`extra_image_6`	string	Optional	URL of reference image 6.
`format`	string	Optional	Output format. Default: `"png"` png

Request Example

json

{
  "model": "magicprompt",
  "image": "https://example.com/room.jpg",
  "prompt": "change the walls to exposed brick and add warm pendant lighting",
  "precise": "40",
  "extra_image_1": "https://example.com/brick-reference.jpg"
}

Response

Success (200)

json

{
  "status": "success",
  "result": {
    "image_url": "https://storage.googleapis.com/...",
    "model": "magicprompt",
    "tokens_used": 1,
    "processing_time_ms": 4500
  }
}

Error

json

{
  "status": "error",
  "error": {
    "code": "INSUFFICIENT_TOKENS",
    "message": "Not enough tokens",
    "required": 1,
    "balance": 0
  }
}

Code Examples

cURL

bash

curl -X POST https://roomai.com/api/v1/generate \
  -H "Authorization: Bearer rai_live_your_api_key" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "magicprompt",
  "image": "https://example.com/room.jpg",
  "prompt": "change the walls to exposed brick and add warm pendant lighting",
  "precise": "40",
  "extra_image_1": "https://example.com/brick-reference.jpg"
}'

JavaScript / Node.js

javascript

const response = await fetch("https://roomai.com/api/v1/generate", {
  method: "POST",
  headers: {
    "Authorization": "Bearer rai_live_your_api_key",
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
  "model": "magicprompt",
  "image": "https://example.com/room.jpg",
  "prompt": "change the walls to exposed brick and add warm pendant lighting",
  "precise": "40",
  "extra_image_1": "https://example.com/brick-reference.jpg"
}),
});

const data = await response.json();

if (data.status === "success") {
  console.log("Image URL:", data.result.image_url);
  console.log("Tokens used:", data.result.tokens_used);
} else {
  console.error("Error:", data.error.message);
}

Python

python

import requests

response = requests.post(
    "https://roomai.com/api/v1/generate",
    headers={
        "Authorization": "Bearer rai_live_your_api_key",
        "Content-Type": "application/json",
    },
    json={
  "model": "magicprompt",
  "image": "https://example.com/room.jpg",
  "prompt": "change the walls to exposed brick and add warm pendant lighting",
  "precise": "40",
  "extra_image_1": "https://example.com/brick-reference.jpg"
},
)

data = response.json()

if data["status"] == "success":
    print("Image URL:", data["result"]["image_url"])
    print("Tokens used:", data["result"]["tokens_used"])
else:
    print("Error:", data["error"]["message"])

Tips

1Be specific in your prompt — mention materials, colors, lighting, and specific elements you want changed.
2Use reference images (extra_image_1 through extra_image_6) to show the AI exactly what style you want.
3Lower "precise" values (10-30) give the AI more creative freedom; higher values (70-90) keep more of the original image.
4Great for quick iterations — describe incremental changes to refine a design.