nodes/TakkunRed-ComfyUI-StepByStep-Sampler
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Add files via upload

Browse Source
This commit is contained in:
Takkun
2026-04-25 01:49:25 +09:00
committed by GitHub
parent f225301f65
commit 32f284d237
1 changed files with 34 additions and 47 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+34 -47
View File
@@ -1,66 +1,53 @@
# ComfyUI-StepByStep-Sampler
# ComfyUI-Metadata-Inspector
A set of custom nodes for ComfyUI designed to analyze and compare the image generation process step-by-step. It quantifies changes in latent variables (using metrics like PSNR/SSIM) and provides a dedicated viewer for intuitive inspection. Furthermore, since sampling automatically stops once image generation converges, image generation can be performed with the optimal number of steps.
A specialized suite of custom nodes for ComfyUI designed to hunt down and visualize generation parameters (prompts, steps, samplers, etc.) embedded in AI-generated images. It works across different formats and platforms (ComfyUI, Automatic1111, Civitai, and more).
[![GitHub license](https://img.shields.io/github.com/TakkunRed/ComfyUI-StepByStep-Sampler)](https://github.com/TakkunRed/ComfyUI-StepByStep-Sampler/blob/main/LICENSE)
[![GitHub license](https://img.shields.io/github/license/TakkunRed/ComfyUI-Metadata-Inspector)](https://github.com/TakkunRed/ComfyUI-Metadata-Inspector/blob/main/LICENSE)
[![ComfyUI](https://img.shields.io/badge/ComfyUI-Custom_Nodes-blue)](https://github.com/comfyanonymous/ComfyUI)
## Core Mission
The primary goal of this tool is to **reveal hidden prompt strings and workflow structures** from AI-generated content. Whether it's a PNG with a complex node graph or a JPEG with a text-based prompt hidden in EXIF data, this inspector brings it to light.
![Node Screenshot](images/workflow_image.png)
## Node description
## Node Descriptions
### 1. Step-by-Step Sampler
Hooks into the generation process to capture intermediate images at specified intervals by VAE encoding the latent state. It can be integrated with `Preview Image` or `Save Image` nodes to display or output the progression as a sequence.
Calculates the amount of change from the previous step using MSE, RMSE, L1, PSNR, or SSIM, allowing you to monitor convergence numerically.
Draws step counts and difference values directly onto the images, making it easy to track the step number and the delta from the preceding frame.
Since sampling automatically stops once image generation converges, image generation can be performed with the optimal number of steps.
Outputs the final result as a standard `LATENT` (similar to KSampler), but can also output the VAE-encoded image via the `LAST_IMAGE` socket.
This extension includes three essential nodes to complete the inspection workflow:
<img src="images/StepByStep_Sampler.png">
### 1. Load Image w/ Name
- **Purpose**: Provides a flexible way to load images.
- **Function**: Unlike the standard loader, it allows you to input a filename via string or absolute path, making it easier to integrate with dynamic file-handling workflows.
### 2. Step-by-Step Player
Visualizes the image list output from the `STEP_IMAGES` socket of the `Step-by-Step Sampler`. Use the slider or the Play button to view the generation process as an animation.
### 2. Extract Metadata (The Core)
- **Purpose**: The "brain" that scans the image file for generation data.
- **Function**:
- **PNG Deep Scan**: Extracts full ComfyUI/A1111 prompt and workflow JSONs.
- **JPEG/EXIF Recovery**: Deep-scans EXIF sub-IFDs to find hidden `UserComment` and `ImageDescription` tags often used by Civitai and Tensor.art.
- **Smart JSON Recovery**: Detects JSON structures even when they are buried inside plain-text metadata.
<img src="images/StepByStepImage.png" width="50%">
### 3. JSON Tree Viewer (Web UI)
- **Purpose**: Provides a human-readable visualization of the raw data.
- **Function**: Transforms the raw metadata string into an interactive, foldable tree view directly on the ComfyUI canvas. It highlights potential prompt areas in orange for quick identification.
### 3. Step-by-Step Comparer
A side-by-side visualization tool for the `STEP_IMAGES` output. It allows you to compare two different steps (e.g., Step 5 vs. Step 20) using an interactive split-screen slider.
<img src="images/comparer.png">
## Key Features
- **Hidden Prompt Extraction**: Recovers data from JPEGs where standard tools fail.
- **Workflow Reconstruction Support**: Outputs raw JSON strings that can be used to understand complex node setups.
- **Multi-Platform Compatibility**: Supports metadata formats from ComfyUI, Automatic1111, and various web-based generators.
## Installation
Manual Install
```
Bash
### Manual Install
```bash
cd ComfyUI/custom_nodes
git clone https://github.com/TakkunRed/ComfyUI-StepByStep-Sampler.git
git clone https://github.com/TakkunRed/ComfyUI-Metadata-Inspector.git
```
## How to Use
### Workflow Integration
* Add the Step-by-Step Sampler node.
* Connect `model`, `positive`, `negative`, `latent_image`, and `vae` to the respective inputs.
* Connect the `STEP_IMAGES` output to the `StepStepPlayer` or `StepStepComparer` node. You can also connect it to `Preview Image` or `Save Image`.
* Turning on `auto_stop` will stop sampling based on the `stop_threshold`.
* Run Queue Prompt as usual, and the generation process will appear in the viewers.
### Viewer Controls
* `Player`: Change steps using the bottom slider or use the "Play" button for auto-playback.
* `Comparer`: Select two steps to compare using the sliders, and drag the vertical bar on the image to reveal the differences between side A and side B.
### Step-by-Step Sampler Settings
* `save_interval`: Determines how often images are saved (e.g., set to 1 to capture every step).
* `show_overlay`: Toggles the on-image display of step numbers and difference metrics.
* `diff_method`: Selects the algorithm for calculating step-to-step changes:
* Metric Characteristics (at convergence):
```
MSE → 0 (Emphasizes large changes)
RMSE → 0 (MSE scaled back to L1 range)
L1 → 0 (Simple, robust against outliers)
PSNR → ∞ (Higher dB means closer to previous step; 40dB+ is a typical convergence target)
SSIM → 1.0 (Structural Similarity Index; 0.99+ is a typical convergence target)
```
* `auto_stop`: Enable/Disable automatic convergence stop
* `stop_threshhold`: Threshold for determining convergence
1. Place Load Image w/ Name.
2. Connect it to Extract Metadata.
3. Connect the metadata_json output to JSON Tree Viewer.
4. Input your image path and run the queue to see the hidden prompt data.
## License
MIT
MIT