VFX

VFX

• The VFX model enables cinematic, high-impact visual effects to be applied to videos or images. It is designed to simulate large-scale visual effects such as explosions, disintegration, levitation, and elemental forces using AI-based compositing and motion synthesis.

Features

• VFX empowers creators with Hollywood-style visual effects through simple prompts or preset selections. Effects are composited onto provided media using spatial, temporal, and visual consistency mechanisms.

Key Features

• AI-Driven Visual FX: Apply dynamic, cinematic effects to static images or videos like explosions, lightning, and tornadoes.
• Preset Effects Library: Supports prebuilt effects including: Building Explosion, Car Explosion etc.
• Spatially-Aware Compositing: Effects are aligned and blended with subjects using AI-inferred object positioning.
• Temporal Control: Timing of effect duration and transitions are automatically calibrated for realism.
• Image Input: Works with static frames and short video clips.

Limitations

• Effect Intensity May Vary: Some presets are highly dependent on subject pose, position, or lighting for best results.
• Single-Subject Priority: Currently optimized for scenes with one or two main visual anchors (e.g., a person or car).
• Static Backgrounds Preferred: Scenes with camera motion or busy backgrounds may reduce realism.
• No Fine-Tuned Control: Users cannot yet fully customize FX positioning, layering, or path animation.

Out-of-Scope Use

• VFX must not be used for

  1. Simulating real-world disaster footage to mislead or cause panic.
  2. Creating visual depictions of harm, violence, or abuse without context or consent.
  3. Modifying imagery of minors in sensitive scenarios.
  4. Misinformation or impersonation campaigns.
  5. Creating scenes that promote hate, extremism, or illegal activities.

Authentication

• For authentication details, please refer to the Authentication Guide.

API Endpoints

Submit Task & Query Result

• cURL

  # Submit Task
  curl --location --request POST 'https://api.muapi.ai/api/v1/generate_wan_ai_effects' \
    --header "Content-Type: application/json" \
    --header "x-api-key: {MUAPIAPP_API_KEY}" \
    --data-raw '{
      "prompt": "a Mercedes bench car",
      "image_url": "https://example.com/car.jpg",
      "name": "Car Explosion",
      "aspect_ratio": "16:9",
      "resolution": "480p",
      "quality": "medium",
      "duration": 5
    }'
  
  # Poll for results
  curl --location --request GET "https://api.muapi.ai/api/v1/predictions/${request_id}/result" \
    --header "x-api-key: {MUAPIAPP_API_KEY}"
  

• Python

  import os
  import requests
  import json
  import time
  
  from dotenv import load_dotenv
  load_dotenv()
  
  def main():
    print("Hello from MuApiApp!")
    API_KEY = os.getenv("MUAPIAPP_API_KEY")
    print(f"API_KEY: {API_KEY}")
  
    url = "https://api.muapi.ai/api/v1/generate_wan_ai_effects"
    headers = {
      "Content-Type": "application/json",
      "x-api-key": f"{API_KEY}",
    }
    payload = {
      "prompt": "a Mercedes bench car",
      "image_url": "https://d3adwkbyhxyrtq.cloudfront.net/ai-images/186/325013990791/fa5a980c-5c3c-42c3-87a3-578976ffd8a6.jpg",
      "name": "Car Explosion",
      "aspect_ratio": "16:9",
      "resolution": "480p",
      "quality": "medium",
      "duration": 5
    }
  
    begin = time.time()
    response = requests.post(url, headers=headers, data=json.dumps(payload))
    if response.status_code == 200:
      result = response.json()
      request_id = result["data"]["request_id"]
      print(f"Task submitted successfully. Request ID: {request_id}")
    else:
      print(f"Error: {response.status_code}, {response.text}")
      return
  
    url = f"https://api.muapi.ai/api/v1/predictions/{request_id}/result"
    headers = {"x-api-key": f"{API_KEY}"}
  
    while True:
      response = requests.get(url, headers=headers)
      if response.status_code == 200:
        result = response.json()
        status = result["data"]["status"]
  
        if status == "completed":
          end = time.time()
          print(f"Task completed in {end - begin} seconds.")
          url = result["video"]["url"]
          print(f"Task completed. URL: {url}")
          break
        elif status == "failed":
          print(f"Task failed: {result.get('error')}")
          break
        else:
          print(f"Task still processing. Status: {status}")
      else:
        print(f"Error: {response.status_code}, {response.text}")
        break
  
      time.sleep(0.5)
  
  if __name__ == "__main__":
    main()
  

• JavaScript

  require('dotenv').config();
  const axios = require('axios');
  
  async function main() {
    console.log("Hello from MuApiApp!");
  
    const API_KEY = process.env.MUAPIAPP_API_KEY;
    console.log(`API_KEY: ${API_KEY}`);
  
    const url = 'https://api.muapi.ai/api/v1/generate_wan_ai_effects';
    const headers = {
      'Content-Type': 'application/json',
      'x-api-key': `${API_KEY}`,
    };
  
    const payload = {
      prompt: 'a Mercedes bench car',
      image_url: 'https://d3adwkbyhxyrtq.cloudfront.net/ai-images/186/325013990791/fa5a980c-5c3c-42c3-87a3-578976ffd8a6.jpg',
      name: '3Car Explosion',
      aspect_ratio: '16:9',
      resolution: '480p',
      quality: 'medium',
      duration: 5,
    };
  
    try {
      const begin = Date.now();
      const response = await axios.post(url, payload, { headers });
  
      if (response.status === 200) {
        const requestId = response.data.data.request_id;
        console.log(`Task submitted successfully. Request ID: ${requestId}`);
  
        const resultUrl = `https://api.muapi.ai/api/v1/predictions/${request_id}/result`;
        const pollHeaders = { x-api-key: `${API_KEY}` };
  
        while (true) {
          const pollResponse = await axios.get(resultUrl, { headers: pollHeaders });
  
          if (pollResponse.status === 200) {
            const status = pollResponse.data.data.status;
  
            if (status === 'completed') {
              const end = Date.now();
              console.log(`Task completed in ${(end - begin) / 1000} seconds.`);
              const videoUrl = pollResponse.data.video.url;
              console.log(`Task completed. URL: ${videoUrl}`);
              break;
            } else if (status === 'failed') {
              console.log(`Task failed: ${pollResponse.data.error}`);
              break;
            } else {
              console.log(`Task still processing. Status: ${status}`);
            }
          } else {
            console.log(`Error: ${pollResponse.status}, ${pollResponse.statusText}`);
            break;
          }
  
          await new Promise(resolve => setTimeout(resolve, 500));
        }
      } else {
        console.log(`Error: ${response.status}, ${response.statusText}`);
      }
    } catch (error) {
      console.error(`Request failed: ${error.message}`);
    }
  }
  
  main();
  

Parameters

Task Submission Parameters

Request Parameters

Parameter	Type	Required	Default	Range	Description
prompt	string	Yes	""	-	The prompt for generating the output.
image_url	string	Yes	""	-	The image for generating the output
name	string	Yes	""	-	Custom name for the task or output file.
aspect_ratio	string	No	16:9	1:1, 9:16, 16:9	Aspect ratio of the output video.
resolution	string	No	480p	480p, 720p	Resolution of the output video.
quality	string	No	medium	medium, high	Controls overall video quality (affects rendering time).
duration	number	No	5	5 ~ 10	Length of the generated video in seconds.

Response Parameters

Parameter	Type	Description
code	integer	HTTP status code (e.g., 200 for success)
message	string	Status message (e.g., “success”)
id	string	Unique identifier for the generation task.
outputs	array of strings	URLs pointing to the generated video output(s).
urls.get	string	API URL to retrieve the result by task ID.
has_nsfw_contents	array of booleans	Indicates whether the output contains NSFW content.
status	string	Status of the task: completed, failed, or processing.
created_at	string (timestamp)	ISO timestamp indicating when the task was created.
error	string	Error message if the task failed; empty if successful.
executionTime	number (milliseconds)	Total time taken to complete the task in milliseconds.
timings.inference	number (milliseconds)	Time taken specifically for model inference in milliseconds.

Result Query Parameters

Result Request Parameters

Parameter	Type	Required	Default	Description
id	string	Yes	-	Task ID

Result Response Parameters

Parameter	Type	Description
code	integer	HTTP status code (e.g., 200 for success)
message	string	Status message (e.g., “success”)
id	string	Unique identifier for the generation task.
outputs	array of strings	URLs pointing to the generated video output(s).
urls.get	string	API URL to retrieve the result by task ID.
has_nsfw_contents	array of booleans	Indicates whether the output contains NSFW content.
status	string	Status of the task: completed, failed, or processing.
created_at	string (timestamp)	ISO timestamp indicating when the task was created.
error	string	Error message if the task failed; empty if successful.
executionTime	number (milliseconds)	Total time taken to complete the task in milliseconds.
timings.inference	number (milliseconds)	Time taken specifically for model inference in milliseconds.