ComfyUI-mnemic-nodes

⛔ Generate Negative Prompt

> [!CAUTION] > This node is highly experimental, and does not produce any useful result right now. It also requires you to download a specially trained model for it. It's just not worth the effort. It's mostly here to share a work in progress project. This node utilizes a GPT-2 text inference model to generate a negative prompt that is supposed to enhance the aspects of the positive prompt. > [!IMPORTANT] > Installation Step: > Download the weights.pt file from the project's Hugging Face repository. > > Place the weights.pt file in the following directory of your ComfyUI setup without renaming it: > > \ComfyUI\customnodes\ComfyUI-mnemic-nodes\nodes\negativeprompt > > The directory should resemble the following structure: > > Installation Directory For additional information, please visit the project's GitHub page.

✂️ String Text Extractor

Compact description unavailable for this node.

✂️ String Text Splitter

Compact description unavailable for this node.

✏️ Literal Bool

Compact description unavailable for this node.

✏️ Literal Float

Compact description unavailable for this node.

✏️ Literal Int

Compact description unavailable for this node.

✏️ Literal String

Compact description unavailable for this node.

✨🌐 Groq ALM API - Translate [EN only]

Compact description unavailable for this node.

✨💬 Groq LLM API

> [!IMPORTANT] > #### 2025-01-12 - Version 1.2.4 > Moved groq API key to a .env instead of a config.ini-file. This will cause existing config setups to break with an update. Apologies for the inconvenience. > #### 2024-09-14 - Version 1.2.0 > This node was renamed to match the new VLM and ALM nodes added. This node makes an API call to groq, and returns the response in text format. You need to manually enter your groq API key into the .env file. A template file is provided and can be renamed to .env for use. Currently, the Groq API can be used for free, with very friendly and generous rate limits. model: Choose from a drop-down one of the available models. The list need to be manually updated when they add additional models. preset: This is a dropdown with a few preset prompts, the user's own presets, or the option to use a fully custom prompt. See examples and presets below. systemmessage: The system message to send to the API. This is only used with the Use [systemmessage] and [userinput] option in the preset list. The other presets provide their own system message. userinput: This is used with the Use [systemmessage] and [userinput], but can also be used with presets. In the system message, just mention the USER to refer to this input field. See the presets for examples. temperature: Controls the randomness of the response. A higher temperature leads to more varied responses. maxtokens: The maximum number of tokens that the model can process in a single response. Limits can be found here. topp: The threshold for the most probable next token to use. Higher values result…

✨📝 Groq ALM API - Transcribe

Query Groq Audio Model to transcribe audio.

✨📷 Groq VLM API

> [!IMPORTANT] > #### 2024-09-27 - Version 1.2.1 > Added new Llama 3.2 vision model to the list, but this model is not yet officially available. Once it is, this should automatically work. This node makes an API call to groq with an attached image and then uses Vision Language Models to return a description of the image, or answer to a question about the image in text format. You need to manually enter your groq API key into the .env file. A template file is provided and can be renamed to .env for use. Currently, the Groq API can be used for free, with very friendly and generous rate limits. Image Size Limit: The maximum allowed size for a request containing an image URL as input is 20MB. Requests larger than this limit will return a 400 error. Request Size Limit (Base64 Enconded Images): The maximum allowed size for a request containing a base64 encoded image is 4MB. Requests larger than this limit will return a 413 error.

🎨 Colorful Starting Image

Compact description unavailable for this node.

🎲 Ideogram 4 Random Prompter

Compact description unavailable for this node.

🎲 Load Random Checkpoint

Compact description unavailable for this node.

🎲 Random Bool

Compact description unavailable for this node.

🎲 Random Color

Compact description unavailable for this node.

🎲 Random Float in Range

Compact description unavailable for this node.

🎲 Random Int in Range

Compact description unavailable for this node.

🎲 Random Seed

Compact description unavailable for this node.

🎲 Random String

Compact description unavailable for this node.

🎵📊 Audio Visualizer

Compact description unavailable for this node.

🏷️ LoRA Loader Prompt Tags

Loads LoRA models using tags in the prompt. The basic syntax is . - The default strength is 1.0 if not specified (e.g., ). - A strength of 0 is treated as 1.0. - If an invalid strength is provided (e.g., non-numeric text), it will also be treated as 1.0. You can provide a separate weight for the CLIP model. - Syntax: - If the CLIP weight is omitted, it defaults to the model's weight. When you use a tag like , the node searches for the best matching LoRA file. If the name is not exact, it uses a scoring system to find the most likely candidate. Here is the order of priority: 1. Exact Match: An exact filename match (e.g., myLora.safetensors) is always preferred. 2. Numbered Versions: It intelligently handles numbered versions (e.g., myLora-1, myLora-2). If you ask for myLora-2, it will prioritize that file. If it can't find it, it will look for other numbered versions or the base myLora file. 3. Prefix Match: It will look for files that start with your requested name (e.g., myLoraAndMore.safetensors). 4. Contains Match: It will look for files that contain your requested name anywhere in the filename. 5. Path Priority: Files located closer to the root of your LoRA folders are given a slight priority boost over files in deep subdirectories. Route the model and clip through the node. Use the output [STRING] to have the prompt without the -tags.

💾 Save Image With Metadata

Save images with Civitai-compatible metadata auto-detected from your workflow. This node is integrated into this pack as an adapted copy of ChronoKnight's original node. Original source: - https://github.com/KChronoKnight/Chrono-Save-for-Civitai - images: Image batch to save. - filenameprefix: Supports strftime, %date:yyyy-MM-dd - hh.mm.ss%, plus %seed and %model. - folder: Output subfolder under ComfyUI output dir, supports strftime and %date:...%. - fileformat: png, jpeg, or webp. - quality: JPEG/WebP quality. - embedworkflow: Embed workflow metadata into saved image. - striploraprompt: Removes and embedding tags from prompt text in metadata output. - positiveoverride (optional): Override positive prompt text. - Captures prompt/negative prompt and LoRAs from executed runtime graph paths (sampler upstream), with graph parsing fallback. - Writes A1111-style parameters metadata. - Models used captures base model + LoRA

💾 Save Text File With Path

This node is adapted and enhanced from the Save Text File node found in the YMC GitHub ymc-node-suite-comfyui pack. The node can now give you a full file path output if you need it, as well as output the file-name as a separate output, in case you need it for something else. > [!IMPORTANT] > #### 2024-06-05 - Version 1.1.1 > The node was severely updated so existing workflows are going to break. I won't do another overhaul like this. > > The new node is more consistent in functionality and more intentional with the inputs and outputs. > > It now handles more edge cases and supports both a prefix, suffix, a dynamic counting with customizable separator before/after the counter in the right circumstances. > > Sorry for any troubles caused.

📁 Get File Path

This node returns the file path of a given file in the \input-folder. It is meant to have a browse-button so you can browse to any file, but it doesn't yet. If you know how to add this, please let me know or do a pull request.

📂 Load Images From Path

Compact description unavailable for this node.

📅 Format Date Time

This node converts date and time formats into literal outputs for you to use. respectsystemlocale toggle functionality: This option allows you to choose between standard SI formats (when disabled) or system-specific locale formats (when enabled) for the %x and %c directives. - Disabled (Default): %x forces YYYY-MM-DD, and %c forces YYYY-MM-DD HH.MM.SS. providing consistent results across different systems. - Enabled: %x and %c use your operating system's locale settings, which may result in formats like MM/DD/YY. %%Y = %Y = Year %%m = %m = Month %%d = %d = Day %%H = %H = Hour %%S = %S = Second %%f = %f = Microsecond %%x = %x = Date (Affected by respectsystemlocale) %%X = %X = Time %%c = %c = Date + Time (Affected by respectsystemlocale) %%p = %p = am/pm %%A = %A = Weekday %%a = %a = Weekday short %%B = %B = Month %%b = %b = Month short %%j = %j = Day of year %%W = %W = Week index (Monday start) %%w = %w = Weekday index (Monday start) %%U = %U = Week index (Sunday start) %%u = %u = Weekday index (Sunday start) %%%% = %% = Prints a literal %%

📐 Resolution Image Size Selector

Creates resolutions and a latent from presets, user presets, or input images. Use the preset drop-down for some image generation standard resolutions. The presetuser input will be created on first launch, and you can edit the config for this in: ComfyUI/customnodes/ComfyUI-mnemic-nodes/nodes/resolutionselector/userresolution.json The custom width and height values are only used with the Custom preset. The multiply scale multiplies the final results and outputs them as multipliedwidth and multipliedheight. This is useful when it comes to figuring out a desired upscale resolution based on variable inputs. The swapwidthandheight option swaps the final width and height. The image input can be used. This has the highest priority. If it is used, we then take the image input width and height. imageminlength ensures that an input image size has at least this length on its shortest side. imagemaxlength ensures that an input image size has at most this length on its longest side. The snaptonearest ensures specific snapping on the final output value. This can be set to 8 or 16 for good divisible sizes, or any value that you need to work with. The batchsize option is used to create multiple output latents when you use tha latent output node.

📝 Wildcard Processor

Compact description unavailable for this node.

📝 Wildcard Processor Advanced

Compact description unavailable for this node.

🔀 Batch Wildcard Upscale Sampler

Resolves wildcards independently for every image, but processes them sequentially

⚡ MNeMiC Nodes · 7 outputs

🔗 String Concat / Append

Concatenates multiple strings together with dynamic input expansion. - Description: Combines multiple strings into one. Features dynamic inputs - new input slots automatically appear as you connect strings, and unused slots are removed when disconnected. - Inputs: - delimiter (optional): Separator to insert between concatenated strings (e.g., , or | ). Leave empty for direct concatenation. - string0, string1, ...: String inputs. More inputs appear automatically as you connect strings. - Outputs: - concatenatedstring: All connected strings joined together. Features: - Dynamic Inputs: The node starts with 2 string inputs. When you connect the last available input, a new empty input automatically appears. Unused inputs at the end are removed when you disconnect. - Empty Handling: Empty inputs are automatically skipped during concatenation. - Delimiter Support: Add any separator between strings, or leave empty for direct concatenation. Example: - Inputs: "Hello", "World", "!" with delimiter - Output: "Hello World !"

🔠 Tiktoken Tokenizer Info

This node takes text as input, and returns a bunch of data from the tiktoken tokenizer. It returns the following values: - tokencount: Total number of tokens - charactercount: Total number of characters - wordcount: Total number of words - splitstring: Tokenized list of strings - splitstringlist: Tokenized list of strings (output as list) - splittokenids: List of token IDs - splittokenidslist: List of token IDs (output as list) - texthash: Text hash - specialtokensused: Special tokens used - specialtokensusedlist: Special tokens used (output as list) - tokenchunkbysize: Returns the input text, split into different strings in a list by the tokenchunksize value. - tokenchunkbysizetoword: Same as above but respects "words" by stripping backwards to the nearest space and splitting the chunk there. - tokenchunkbysizetosection: Same as above, but strips backwards to the nearest newline, period or comma.

🖼️ Download Image from URL

This node downloads an image from an URL and lets you use it. It also outputs the Width/Height of the image. By default, it will save the image to the /input directory. Clear the savepath line to prevent saving the image (it will still be saved in the TEMP-folder). If you enter a name in the savefilenameoverride section, the file will be saved with this name. You can enter or ignore the file extension. If you enter one, it will rename the file to the chosen extension without converting the image. Supported image formats: JPG, JPEG, PNG, WEBP. Does not support saving with transparency. > [!IMPORTANT] > #### 2024-09-14 - Version 1.2.0 > This node was renamed in the code to match the functionality. This may break existing nodes.

🖼️ Load Image Advanced

Compact description unavailable for this node.

🖼️ Load Image Temporarily

Compact description unavailable for this node.

🖼️+📝 Load Text-Image Pair (Single)

This node loads only one specific image-text pair from a folder. It is designed to be extremely fast and lightweight. Key Feature: No Caching This node does not cache any data. It reads the requested files directly from the disk on every run. Use this when you only need a single item from a dataset and don't want to spend time pre-loading the entire set. - seed: The index of the pair to load. - folderpath: Path to a folder containing image and text files with matching names. - imageinput: (Optional) A single image or a batch of images. Overrides folderpath. - textinput: (Optional) A single string or a list of strings. Overrides folderpath. - textformatextension: (Optional) The extension for text files (e.g., txt, caption). Default is txt. - imagesingle: The single image selected by the seed. - stringsingle: The text paired with the selected image. - imagepathsingle: The full absolute path of the selected image. - imagefilenamesingle: The filename (without extension) of the selected image. - totalcount: The total number of pairs found in the dataset.

🖼️+📝 Load Text-Image Pairs (List)

This node loads an entire dataset of image and text pairs from a folder. It is designed for performance when you need to iterate through or use the full list of items. Key Feature: Caching The first time this node is run on a folder, it loads all images, processes them into a batched tensor, and stores the result in memory. Subsequent runs are nearly instantaneous as they use the cached data. > [!IMPORTANT] > To cycle through each input entry one by one in the list, you'll want to set the seed input to increment. - seed: The starting index for selecting pairs. Set to increment to cycle through the dataset. - folderpath: Path to a folder containing image and text files with matching names. - forcereload: A checkbox to force the node to discard the cache and reload all data from the disk. - limitcount: (Optional) The maximum number of pairs to return in the list outputs. 0 returns all pairs. - textformatextension: (Optional) The extension for text files (e.g., txt, caption). Default is txt. - imagesingle: The single image selected by the seed. - stringsingle: The text paired with the selected image. - imagepathsingle: The full absolute path of the selected image. - imagefilenamesingle: The filename (without extension) of the selected image. - imagelist: A list of all images in the dataset. - stringlist: A list of all text strings in the dataset. - imagepathlist: A list of full absolute paths for all images. - imagefilenamelist: A list of filenames for all images. - totalcount: The total number of pairs found in the dataset.

🖼️📊 Metadata Extractor (List)

Extracts metadata from a list of images or a folder of images. The input images are loaded via a path. Either directly to an image file, or to a folder with multiple images. The metadata should work on both jpg and png images, and it will do what it can to extract from A1111/Forge/ComfyUI image metadata. > [!IMPORTANT] > To cycle through each input entry one by one in the list, you'll want to set the seed input to increment. > [!WARNING] > Be mindful of the maxfilecount setting when working with large folders. Loading a large number of images can consume a significant amount of memory and slow down your workflow. It is recommended to set a reasonable limit to avoid performance issues. Without a maxfilecount limit, things will take a long time to load each time and comfy will get super laggy if you load a folder with hundreds or thousands of images. Each will be loaded into memory at the same time. Use the Metadata Extractor (Single) if you only need one image at a time! - Inputs: - seed: An integer that sets the starting index for selecting image-metadata pairs. Use the increment option in workflow settings to cycle through pairs sequentially. - inputpath: A string specifying the path to a folder containing images or a single image file. Only used if imageinput is not provided. - imageinput: A single image or a batch/list of images. Takes priority over inputpath. - filterparams: A comma-separated string of case-insensitive keys (e.g., steps, sampler, seed) to extract specific metadata parameters. - maxfilecount: An integer limiting the total number of image-metadata pairs…

🖼️📊 Metadata Extractor (Single)

Extracts metadata from a single image selected from an input source. The input images are loaded via a path, either directly to an image file or to a folder with multiple images. The metadata should work on both jpg and png images, and it will do what it can to extract from A1111/Forge/ComfyUI image metadata. > [!IMPORTANT] > To cycle through each input entry one by one in the list, you'll want to set the seed input to increment. - Inputs: - seed: An integer that sets the starting index for selecting image-metadata pairs. Use the increment option in workflow settings to cycle through pairs sequentially. - inputpath: A string specifying the path to a folder containing images or a single image file. Only used if imageinput is not provided. - imageinput: A single image or a batch/list of images. Takes priority over inputpath. - filterparams: A comma-separated string of case-insensitive keys (e.g., steps, sampler, seed) to extract specific metadata parameters. - Outputs: - image: The single image selected based on the seed index from the input list. - positiveprompt: The positive prompt extracted from the selected image's metadata. - negativeprompt: The negative prompt extracted from the selected image's metadata. - parsedparamsjson: All parsed metadata parameters from the selected image, returned as a JSON string. - filteredparamslist: Specific metadata values requested in filterparams from the selected image, returned as a multi-line string. - rawmetadatajson: The complete, unprocessed metadata of the selected image, formatted as a JSON string. For extracted params, you can…

🧩 Ideogram 4 Prompt Builder w. String Inputs

Compact description unavailable for this node.

🧹 String Cleaning

Compact description unavailable for this node.

AudioVisualizer

Creates stunning audio visualizations from audio files with multiple visualizer styles and output modes. Supports particles, waveforms, and geometric patterns with automatic state reset for clean animation starts.

⚡ MNeMiC Nodes · 3 outputs

ColorfulStartingImage

Creates a highly customizable, colorful image with random patterns and shapes, useful as a starting point for image generation.

⚡ MNeMiC Nodes · 0 inputs · 21 parameters · 2 outputs

Ideogram4PromptBuilder

Visual prompt builder for Ideogram 4's structured JSON caption format.

⚡ MNeMiC Nodes · 2 inputs · 12 parameters · 5 outputs

Ideogram4RandomPrompter

Experimental random prompt generator for Ideogram 4's structured JSON caption format.

⚡ MNeMiC Nodes · 0 inputs · 27 parameters · 5 outputs

ImageSaveWithMetadata

Save images with Civitai-compatible metadata. Auto-extracts model,

⚡ MNeMiC Nodes · 2 inputs · 6 parameters · 0 outputs

LiteralBool

A simple boolean literal input. Toggle between True and False.

⚡ MNeMiC Nodes · 0 inputs · 1 parameter · 1 output

LiteralFloat

A simple floating-point literal input. Enter any decimal number.

⚡ MNeMiC Nodes · 0 inputs · 1 parameter · 1 output

LiteralInt

A simple integer literal input. Enter any whole number.

⚡ MNeMiC Nodes · 0 inputs · 1 parameter · 1 output

LiteralString

A simple string literal input. Enter any text, supports multiple lines.

⚡ MNeMiC Nodes · 0 inputs · 1 parameter · 1 output

LoadImageAdvanced

Loads an image and extracts its file path and positive prompt from metadata.

⚡ MNeMiC Nodes · 0 inputs · 1 parameter · 6 outputs

LoadImagesFromPath

Loads a single image from a directory, allowing sequential iteration through the folder.

⚡ MNeMiC Nodes · 0 inputs · 2 parameters · 5 outputs

LoadImageTemporarily

Loads an image and stores it in ComfyUI /temp-folder instead of the /input folder.

⚡ MNeMiC Nodes · 4 outputs

LoadRandomCheckpoint

Load checkpoints from a flexible list with repeat control. Supports fuzzy name matching, file paths, and directories. Perfect for batch processing with varied model selection.

⚡ MNeMiC Nodes · 0 inputs · 4 parameters · 4 outputs

LoraTagLoader

Loads LoRA tags from the provided input string (usually the prompt) and applies them to the model without needing one or multiple LoRA Loader nodes

3 inputs · 0 parameters · 3 outputs

PromptPropertyExtractor

The Prompt Property Extractor node is a tool designed to parse a string (like a prompt) and extract various workflow properties such as model checkpoints, VAEs, LoRAs, sampler settings, and even generate latent tensors. It allows you to define a large part of your workflow within a single text block, making it a useful tool for randomizing more parts of your generations than just a wildcard. The node accepts a primary input string and a set of default values for all supported properties. If a property is found in the input string, it will override the default value. - inputstring: The main text string to parse for properties. - model, clip, vae: Default model, CLIP, and VAE to use if not specified in the string. - cfg, steps, sampler, etc.: Default values for all standard sampler and image properties. It is important to understand which model/CLIP/VAE is used when multiple sources are available (Input pins vs. Tags). Priority Order (Highest to Lowest): 1. Specific Tag: or always takes the highest priority. 2. Checkpoint Tag: If a tag is present: Model: Always loaded from the checkpoint. CLIP: Loaded from checkpoint IF loadclipfromcheckpoint is True. VAE: Loaded from checkpoint IF loadvaefromcheckpoint is True. 3. Input Pin: The model, clip, and vae inputs are used if they are not overridden by the above. Examples: loadclipfromcheckpoint = True + Input CLIP + tag: The Checkpoint's CLIP is used (Input is ignored). loadclipfromcheckpoint = False + Input CLIP + tag: The Input CLIP is used. loadclipfromcheckpoint = True + Input CLIP + (NO tag): The Input CLIP is used (Setting i…

⚡ MNeMiC Nodes · 19 outputs

RandomBool

Generates a random boolean value (True/False). Adjust probability to weight the randomness. Use seed for reproducibility or -1 for random results.

⚡ MNeMiC Nodes · 0 inputs · 2 parameters · 1 output

RandomColor

Generates a random RGB color. Returns both hex format (#RRGGBB) and individual RGB values (0-255). Use seed for reproducibility or -1 for random results.

⚡ MNeMiC Nodes · 0 inputs · 1 parameter · 4 outputs

RandomFloatInRange

Generates a random floating-point number within the specified range. Optionally round to specified decimal places. Use seed for reproducibility or -1 for random results.

⚡ MNeMiC Nodes · 0 inputs · 4 parameters · 1 output

RandomIntInRange

Generates a random integer within the specified range (min to max, inclusive). Use seed for reproducibility or -1 for random results.

⚡ MNeMiC Nodes · 0 inputs · 3 parameters · 1 output

RandomSeed

Generates a random seed value (0 to 2^64-1). Useful for feeding into other nodes that accept seed inputs. Regenerates on each execution.

⚡ MNeMiC Nodes · 0 inputs · 0 parameters · 1 output

RandomString

Randomly selects one string from a newline-separated list of options. Empty lines are ignored. Use seed for reproducibility or -1 for random results.

⚡ MNeMiC Nodes · 0 inputs · 2 parameters · 1 output

ResolutionSelector

Flexible resolution selector with presets, image input, min/max lengths, snapping, swapping and resolution multiplication.\n\nPriority order for resolution selection:\n1. Input Image\n2. User Preset\n3. Preset\n4. Custom Values (when preset is set to 'Custom')

⚡ MNeMiC Nodes · 1 input · 11 parameters · 3 outputs

StringCleaning

Cleans up the input text based on various stripping options.

⚡ MNeMiC Nodes · 1 input · 18 parameters · 1 output

StringConcat

Concatenates multiple strings together with an optional delimiter. Connect strings to the inputs - more inputs automatically appear as needed. Empty inputs are skipped.

⚡ MNeMiC Nodes · 2 inputs · 1 parameter · 1 output

StringTextExtractor

Extracts the first occurrence of text between a pair of characters (delimiters).

⚡ MNeMiC Nodes · 0 inputs · 2 parameters · 3 outputs

StringTextSplitter

Splits a string by the first occurrence of a delimiter.\n\nExample:\n- input_string: part1|part2|part3\n- delimiter: |\n- first_chunk: part1\n- remainder: part2|part3

⚡ MNeMiC Nodes · 0 inputs · 4 parameters · 4 outputs

TiktokenTokenizer

Tokenizes input text and returns various tokenization details, including token count, special tokens used, and more.

⚡ MNeMiC Nodes · 0 inputs · 2 parameters

WildcardProcessor

A text processor that replaces wildcards with dynamic content from files or inline lists.

⚡ MNeMiC Nodes · 0 inputs · 3 parameters · 1 output