Elevate Your Image Inpainting with the Multi-Controlnet & Consistency Decoder Actions

22 Apr 2025
Elevate Your Image Inpainting with the Multi-Controlnet & Consistency Decoder Actions

In the ever-evolving realm of image processing, the usamaehsan/multi-controlnet-x-consistency-decoder-x-realestic-vision-v5 spec offers a powerful toolset for developers looking to enhance their image manipulation capabilities. With the Perform Multi-Controlnet Inpainting action, you can leverage advanced models to achieve remarkable results in image inpainting. This action allows for intricate adjustments, enabling you to refine image quality through techniques like prompt-weight adjustments and tile conditioning.

Prerequisites

To get started with the Cognitive Actions, you'll need an API key for the Cognitive Actions platform. Authentication typically involves passing this API key in the headers of your requests. Ensure you have your key ready as you integrate these actions into your applications.

Cognitive Actions Overview

Perform Multi-Controlnet Inpainting

The Perform Multi-Controlnet Inpainting action utilizes advanced image processing techniques to fill in missing parts of images intelligently. By harnessing both Multi-Controlnet and consistency-decoder models, this action allows for intricate adjustments to image qualities such as prompt weighting, tile conditioning, and low-resolution fixes.

Input

The input schema for this action is designed to accommodate a variety of parameters, enabling you to customize the inpainting process:

  • prompt (required): The text prompt used for image generation. Use +++ to increase the weight of specific words.
  • seed (optional): An integer to set the random number generation starting point for reproducibility.
  • guessMode (optional): A boolean to determine if the model should attempt to recognize content without prompts.
  • lowResFix (optional): A boolean to enable post-generation processing for improved resolution.
  • maskImage (optional): A URI for the mask image employed in inpainting.
  • scheduler (optional): Select a scheduling algorithm for image generation. Default is DDIM.
  • tiledImage (optional): A URI for the control image used for tile ControlNet.
  • lineArtImage (optional): A URI for the control image used for lineart ControlNet.
  • maximumWidth (optional): Maximum width of the generated image in pixels (default is 512).
  • guidanceScale (optional): Intensity of classifier-free guidance during generation (default is 7).
  • maximumHeight (optional): Maximum height of the generated image in pixels (default is 512).
  • negativePrompt (optional): Text prompts for undesired features.
  • numberOfOutputs (optional): The number of images to generate (default is 1).
  • ... more fields available in the schema

Example Input:

{
  "prompt": "underwater++ world",
  "guessMode": false,
  "scheduler": "DDIM",
  "lineArtImage": "https://replicate.delivery/pbxt/Jqzz9g4kPNdw0IyCqEGIXBHCr33Ai9agnbnZGdNVUUKoJqt3/955ad4a4680988283115264b5b45c211.jpg",
  "maximumWidth": 512,
  "guidanceScale": 7,
  "maximumHeight": 512,
  "negativePrompt": "Longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality",
  "numberOfOutputs": 1,
  "disableSafetyCheck": true,
  "estimatedTimeArrival": 0,
  "tileConditioningScale": 1,
  "numberOfInferenceSteps": 20,
  "lineartConditioningScale": 1,
  "brightnessConditioningScale": 1,
  "inpaintingConditioningScale": 1
}

Output

The output from this action typically returns a URL to the generated image after processing. Here’s an example of what you might receive:

Example Output:

[
  "https://assets.cognitiveactions.com/invocations/7a13e415-3c1d-4ec4-ae05-086466d5a960/3586dc37-c905-47e1-8f89-682b6b114694.png"
]

Conceptual Usage Example (Python)

Below is a conceptual Python code snippet demonstrating how you might call this action using a hypothetical Cognitive Actions execution endpoint:

import requests
import json

# Replace with your Cognitive Actions API key and endpoint
COGNITIVE_ACTIONS_API_KEY = "YOUR_COGNITIVE_ACTIONS_API_KEY"
COGNITIVE_ACTIONS_EXECUTE_URL = "https://api.cognitiveactions.com/actions/execute" # Hypothetical endpoint

action_id = "adeace85-cceb-4fb0-ad1a-0d6dff06a361" # Action ID for Perform Multi-Controlnet Inpainting

# Construct the input payload based on the action's requirements
payload = {
    "prompt": "underwater++ world",
    "guessMode": False,
    "scheduler": "DDIM",
    "lineArtImage": "https://replicate.delivery/pbxt/Jqzz9g4kPNdw0IyCqEGIXBHCr33Ai9agnbnZGdNVUUKoJqt3/955ad4a4680988283115264b5b45c211.jpg",
    "maximumWidth": 512,
    "guidanceScale": 7,
    "maximumHeight": 512,
    "negativePrompt": "Longbody, lowres, bad anatomy, bad hands, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality",
    "numberOfOutputs": 1,
    "disableSafetyCheck": True,
    "estimatedTimeArrival": 0,
    "tileConditioningScale": 1,
    "numberOfInferenceSteps": 20,
    "lineartConditioningScale": 1,
    "brightnessConditioningScale": 1,
    "inpaintingConditioningScale": 1
}

headers = {
    "Authorization": f"Bearer {COGNITIVE_ACTIONS_API_KEY}",
    "Content-Type": "application/json"
}

try:
    response = requests.post(
        COGNITIVE_ACTIONS_EXECUTE_URL,
        headers=headers,
        json={"action_id": action_id, "inputs": payload} # Hypothetical structure
    )
    response.raise_for_status() # Raise an exception for bad status codes (4xx or 5xx)

    result = response.json()
    print("Action executed successfully:")
    print(json.dumps(result, indent=2))

except requests.exceptions.RequestException as e:
    print(f"Error executing action {action_id}: {e}")
    if e.response is not None:
        print(f"Response status: {e.response.status_code}")
        try:
            print(f"Response body: {e.response.json()}")
        except json.JSONDecodeError:
            print(f"Response body: {e.response.text}")

In this snippet, you'll notice how the action ID and input payload are structured for the API call. Adjust the endpoint URL and request structure as needed for your implementation.

Conclusion

The Perform Multi-Controlnet Inpainting action provides developers with an advanced mechanism for image inpainting, allowing for fine-tuning of image quality through various parameters. By integrating these Cognitive Actions into your applications, you can significantly enhance your image processing workflows. Explore the possibilities and push the boundaries of what you can achieve with your image content!

← Back to All Articles