ComfyUI Extension

Dehypnotic

Creative suite for ComfyUI: Aspect Ratio, Numbered Text, Save-nodes (MP3 Audio, Images, and Video), Range To String, and Set/Get.

By Dehypnotic·Created 30 days ago·Updated 8 days ago· 4

Dehypnotic/ComfyUI-Dehypnotic

Nodes—

On cloudLocal install

Stars4

Updated8 days ago

Readme

Important: This is my only repository now, so delete any old folders (comfyui-aspect-ratio-advanced, comfyui-numbered-text, comfyui-range-to-string, or comfyui-dehypnotic-save-nodes) you may have in ComfyUI\custom_nodes to avoid conflicts.

ComfyUI-Dehypnotic Custom Nodes

A suite of feature-rich, high-performance custom nodes for ComfyUI. This repository provides nodes for image, video, audio saving, dynamic aspect ratio calculation, wireless data rerouting, and text generation.

Nodes Overview
Security and External Save Paths
Installation
License

Nodes Overview

🧘 AspectRatio (Dehypnotic)

Class Name: AspectRatio / dehypnotic_AspectRatio
Category: Dehypnotic/📐 Aspect Ratio

A flexible aspect ratio and resolution generator node with interactive frontend controls, visual aspect ratio presets, image reference scaling, grid snapping, and VAE encoding support.

Key Features:

Preset & Custom Ratio Modes: Choose standard presets (1:1, 16:9, 4:3, etc.) or set custom width/height ratios and dimensions.
Reference Image Scaling: Connect an optional input IMAGE to scale it to the calculated aspect ratio/dimensions using Lanczos, Bicubic, Bilinear, Area, or Nearest Exact interpolation.
Grid Snapping: Automatically rounds dimensions to the nearest multiple of 8, 16, 32, or 64 (ideal for diffusion models like SD1.5, SDXL, FLUX, etc.).
Optional VAE Encoding: Connect a VAE model to directly encode the scaled image into a LATENT representation.

Inputs & Outputs:

🧘 Set Dehypnotic & Get Dehypnotic

Class Name: DehypnoticSetNode, DehypnoticGetNode
Category: Dehypnotic/🔀 Wireless Links

Wireless routing nodes designed to keep your ComfyUI node graphs organized, readable, and free of crossing connection wires ("spaghetti cables").

Key Features:

Universal Data Support: Works with any ComfyUI data type (IMAGE, LATENT, MODEL, CLIP, CONDITIONING, INT, FLOAT, STRING, custom types, etc.).
Virtual Node Architecture: Operates on the frontend with zero backend execution overhead. ComfyUI automatically resolves connections directly from the source during graph execution.
Passthrough Output: Set Dehypnotic outputs the connected value so you can chain nodes without extra splits.

How to Use:

Attach any output to a Set Dehypnotic node and enter a variable name.
Place a Get Dehypnotic node anywhere in your graph, select the variable name from the dropdown, and connect its output to your target node.

🧘 NumberedText (Dehypnotic)

Class Name: NumberedText / dehypnotic_NumberedText
Category: Dehypnotic/📝 Text Utils

A prompt management and text block organizer node. Allows writing multi-line text entries with interactive checkbox toggles to selectively combine prompts. Now features a robust preset saving system.

Key Features:

Interactive Numbered Blocks: Create new numbered entries using Enter. Create multi-line sub-texts within the same item using Shift + Enter.
Selective Output: Toggle checkboxes ([x] / [ ]) next to numbered items. Only checked text items are combined and output.
Preset Management: Save and load your text lists as .json files directly from the UI. A dropdown picker lets you easily load, update, or delete saved files. Automatically creates and stores files under ComfyUI/user/Dehypnotic/numbered_text/.
Custom Delimiters: Joins active text blocks using any custom separator (e.g., , , \n, |).
Swap & Clone Operations: Quick UI shortcuts to swap the content of any two numbered items, or clone an item into a target position (inserting it like Enter does, shifting all subsequent item numbers down by 1). Steppers start at 1 by default and their values persist after execution. The target index's max value is set to 1 more than the current sequence size, enabling cloning directly to a new item. Operation on non-existent entries behaves as if they were empty.
Selection Modes (Single / Multi): Dedicated toggle button located on the bottom control row (left side of Swap/Clone line) to switch between Single mode (at most 1 item checked at a time; checking a new item unchecks the previous one, and activating Single mode clears extra checkmarks if multiple were selected; disables "Check All") and Multi mode (multiple items checked simultaneously).
Quick Operations: Dedicated equal-width buttons for "New" (creates a new empty text at the next available sequence number), "Delete" (deletes checked lines), "Copy" (copies checked lines to clipboard), "Paste" (pastes clipboard text directly as a new sequence item), "Check All" (disabled in Single mode), and "Uncheck All". Includes informative tooltips on hover for all buttons, steppers, swap, clone, and separator controls.

Inputs & Outputs:

🧘 Text (Dehypnotic)

Class Name: Text
Category: Dehypnotic/📝 Text Utils

A pure text block node equipped with a plain multi-line text editor, built-in file saving/loading (.txt files), and clipboard convenience buttons.

Key Features:

Plain Text Editor: Just a classic text field without numbers or checkboxes.
Preset Management: Load and save plain text as .txt files directly from the UI dropdown. Automatically creates and stores files under ComfyUI/user/Dehypnotic/text/.
Quick Operations & Input Toggle: Dedicated "Input ON / Input OFF" toggle button (placed next to Copy) to easily enable or disable incoming text from text_in without disconnecting wires. Also includes "Copy", "Paste", and "Clear" buttons at the bottom.
Dynamic Text Replacing: An optional text_in input (displayed as text on the node). When Input ON is active (default), text provided from another node will replace the text in this node upon execution. When Input OFF is active, incoming text is ignored and the text in the editor remains unchanged.

Inputs & Outputs:

🧘 RangeToString (Dehypnotic)

Class Name: RangeToString / dehypnotic_RangeToString
Category: Dehypnotic/📝 Text Utils

Generates a formatted string representing a numerical sequence. Useful for batching, prompt scheduling, frame indices, or automated parameter sweeps.

Key Features:

Supports forward and reverse numerical ranges (positive or negative steps).
Configurable range boundary rules (inclusive or exclusive).
Customizable delimiter string.

Inputs & Outputs:

🧘 Save Audio (Dehypnotic)

Class Name: SaveAudioMP3
Category: Dehypnotic/💾 IO

A specialized, multi-format audio output node with dynamic UI and built-in playback. Seamlessly encode audio to MP3, WAV, FLAC, or Opus.

Key Features:

Multi-Format Support: Save audio as MP3, WAV (16-bit or 24-bit), FLAC, or Opus.
Dynamic Interface: The node's widget layout automatically adapts to show only the settings relevant to the selected audio format, keeping your graph clean.
Inline Audio Player: Preview your saved audio directly inside the ComfyUI node. Includes play/pause, seek bar, volume control, and an autoplay toggle.
Preview Only Mode: Toggle preview_only (on/off) to audition/preview audio streams in the built-in player without saving output files to disk (off by default).
Format-Specific Tuning:
- MP3: variable (VBR), constant (CBR), and average (ABR) encoding across low, medium, and high quality.
- WAV/FLAC: Choose target sample rate and 16-bit or 24-bit depth.
- FLAC: Configurable compression level.
- Opus: Configurable bitrate, application mode (audio or voip), and VBR toggle.
Audio Format Auto-Normalization: Handles mono/stereo inputs, arbitrary sample rates, numpy arrays, and PyTorch audio tensors seamlessly via FFmpeg.
Dynamic Path Expansion: Supports date placeholders (%Y-%m-%d), timestamp variables, environment variables, and unique IDs in output paths.

Inputs & Outputs:

🧘 Save Images (Dehypnotic)

Class Name: SaveImagesDehypnotic
Category: Dehypnotic/💾 IO

An advanced multi-format image saving node featuring sequence numbering, date-based folder grouping, image optimization, and workflow metadata embedding.

Key Features:

Interactive Thumbnail Gallery: Displays a scrollable history of recently saved images directly on the node with an elegant, space-saving UI.
Fullscreen Image Viewer: Click any thumbnail in the gallery to open a full-screen, high-resolution preview overlay.
Toggleable Preview: Hide or show the gallery instantly via a custom toggle button to free up canvas space.
Smart UI Layout: Configurable parameters like number_padding and number_start are automatically moved to the node's right-click Properties panel, keeping the main node body compact.
Multi-Format Export: Supports PNG, JPG/JPEG, WEBP, GIF, BMP, and TIFF formats.
Workflow Metadata Embedding: Embeds full ComfyUI workflow metadata into PNG (via tEXt chunks) and WebP images (via XMP metadata).
Sequential File Naming: Automatic file index incrementing with customizable zero-padding (e.g. 0001, 0002) and prefix/delimiter settings.
Quality & Compression Controls: Configurable image quality percentage, WEBP lossless encoding option, PNG/JPG image optimization, and custom DPI metadata.
Compactness: number_padding, number_start, and dpi moved to the properties panel for a more compact node.

Inputs & Outputs:

🧘 FrameSave (Dehypnotic)

Class Name: FrameSave / FrameSaveDehypnotic
Category: Dehypnotic/💾 IO

An interactive frame selection, inspection, and temporary saving node designed for working with image sequences and animation batches. Automatically clears its temporary folder on every execution, filters frame ranges with custom stepper controls, and provides an interactive full-brightness gallery with full-resolution zoom.

Key Features:

Automatic Temp Folder Clearance: Empties old temporary files from temp/dehypnotic_frame_save automatically on every run before saving the current filtered frame batch.
Range & Step Filtering: Filter visible and saved frames using 1-indexed Start frame, End frame (0 = all remaining frames), and Step interval controls.
Custom [-] & [+] Steppers: Clean stepper buttons around Start frame, End frame, and Step fields for intuitive value adjustments.
Live Upstream Frame Count: Automatically traces input graph links recursively (past Set/Get wireless nodes, AspectRatio, and processing nodes) to detect and display the total frame count of connected video/image sources (Total: N) live before workflow execution.
In-Browser Directory Picker Modal: Built-in interactive folder browser modal (📁) that allows visual directory and drive navigation (C:\, D:\, Output, etc.) with zero browser upload security prompts.
Full-Brightness Gallery & Fullscreen Zoom: Scrollable 500px thumbnail gallery displaying filtered frames at 100% brightness with mint-green (#34d399) selection borders, double-click fullscreen zoom overlay, Shift + Click range selection, and instant "Select All" / "Deselect All" buttons. Default selection is set to none on first view.
Security & Whitelist Validation: Ensures saving to chosen target locations strictly obeys ComfyUI output path restrictions or custom allowed paths defined in dehypnotic_save_allowed_paths.json.

Inputs & Outputs:

🧘 Load Video (Dehypnotic)

Class Name: LoadVideoDehypnotic
Category: Dehypnotic/IO

A streamlined video loading node with an interactive frontend, built-in file uploading, and drag-and-drop support. It extracts frames and audio from video files seamlessly.

Key Features:

Built-in Upload & Drag-and-Drop: Easily upload videos directly into ComfyUI's input folder using the "Upload video" button, or simply drag and drop a video file onto the node.
Dynamic Video Player: Features a compact, native video player at the bottom of the node to preview the selected video (plays on mouseover).
Format Support: Automatically lists known video formats (mp4, mkv, webm, avi, mov, gif) from your input directory, sorted by newest first.
Frame & Audio Extraction: Outputs high-quality IMAGE tensors (via OpenCV) and extracts audio tracks natively into ComfyUI's standard AUDIO format (via FFmpeg).

Inputs & Outputs:

🧘 Save Video (Dehypnotic)

Class Name: SaveVideoDehypnotic
Category: Dehypnotic/💾 IO

A comprehensive video renderer and frame exporter node leveraging bundled imageio-ffmpeg for high quality video generation with optional audio multiplexing.

Key Features:

Flexible Inputs: Connect either image frame batches (IMAGE) or video inputs (VIDEO file path, video tensor, or dictionary).
Automatic Video & Audio Extraction: When connecting a VIDEO input, frame sequences and audio tracks (if present and no external AUDIO is attached) are automatically extracted and passed through the encoding pipeline.
Multiple Containers & Professional Codecs:
- Containers: mp4, mkv, webm, mov.
- Codecs: H.264 (libx264), H.265/HEVC (libx265), VP9 (libvpx-vp9), AV1 (libaom-av1), ProRes 422 HQ (prores_ks), DNxHR HQ (dnxhr_hq).
Audio Integration & Single-Frame Looping: Attach mono or stereo AUDIO. If a single image frame and an audio track are provided, the node automatically loops the frame for the full duration of the audio.
Quality & Performance Tuning: CRF (Constant Rate Factor) quality control, encoder speed presets (ultrafast to veryslow), and optional frame preview rendering in the node output.
Preview Only Mode: Toggle preview_only (off/on) to audition video previews in the node player without saving video files to disk (off by default).

Key Inputs & Outputs:

🧘 SonicColor Noise (Dehypnotic)

Class Name: SonicColor
Category: Dehypnotic/Audio

A feature-rich noise generator and audio processor with an embedded dark GUI interface. Features 5 noise colors (White, Pink, Brown, Blue, Violet), a resonant filter with 5 modes, a 10-Band graphic equalizer, ADSR amplitude envelope shaping, real-time WebAudio live playback preview, user preset management, and non-looping audio mixing.

Key Features:

5 Noise Color Mixing: Blend White, Pink (Paul Kellet filter), Brown (Leaky integrator), Blue, and Violet noise with individual gain controls.
Embedded Custom GUI: 3-tab dark-themed UI (Color Mix, Filter & Env, 10-Band EQ) with a fixed height and smooth internal scrollbar.
Real-Time Live WebAudio Preview: Interactive ▶️ Play/Stop preview button allowing you to hear adjustments to noise colors, filters, ADSR, and EQ in real time.
Real-Time Media Duration Detection: Automatically detects and populates media duration live from connected Load Audio or Load Video nodes, rounded up to whole seconds.
Non-Looping External Audio Mix: Mixes connected input audio once (padded with silence if shorter than target duration) with adjustable input volume.
Resonant Filter & 10-Band Equalizer: Built-in Lowpass, Highpass, Bandpass, and Notch filters plus a 10-Band graphic EQ (50Hz to 16kHz) for precise sound design.
Preset System: Easily save, load, update, and delete custom user presets directly from the GUI (saved safely in ComfyUI/user/Dehypnotic/sonic_color/).

Inputs & Outputs:

🧘 BrainwaveSync (Dehypnotic)

Class Name: BrainwaveSync
Category: Dehypnotic/🔊 Audio

A compact, interactive brainwave entrainment audio generator node featuring a dynamic HTML5 canvas graph editor, preset manager, Web Audio live preview, and high-quality audio output generation (Isochronic pulses or Binaural beats).

Key Features:

Interactive Frequency-over-Time Graph: Click, double-click, or Shift-click to add points, drag points to adjust beat frequencies, and right-click to delete. Automatically resizes with the node boundaries.
Isochronic & Binaural Modes:
- Isochronic: Synthesizes mono audio with carrier frequency modulated by triangular pulse envelopes.
- Binaural: Synthesizes stereo audio with dual carrier frequencies ($f_{left} = f_{carrier}$, $f_{right} = f_{carrier} + f_{beat}$).
Web Audio Live Preview: Integrated ▶️ Play / ⏹ Stop browser audio engine to preview entrainment sessions live.
Preset Management: Load, save, duplicate, and delete custom sessions saved in ComfyUI/user/Dehypnotic/brainwave_sync/.
Grid Snapping: Time snapping (Off, 10s, 30s, 1m, 5m) and Hz snapping (Off, 0.1Hz, 0.5Hz, 1.0Hz) for precise curve editing.
Audio Output Pipeline: Direct AUDIO dict output (waveform tensor + sample_rate) ready for Save Audio (Dehypnotic) or standard ComfyUI audio nodes.

Inputs & Outputs:

Security and external save paths (ComfyUI Manager compliant)

By default, saving is allowed under ComfyUI’s output/ directory.
To allow external locations (e.g., other drives), create a local text-file next to this node named dehypnotic_save_allowed_paths.json containing for example:
```
{ "allowed_roots": ["D:/AudioExports", "E:/TeamShare/Audio"] }
```

Whitelist behavior and safety

The node refuses to write outside ComfyUI/output/ unless the path is under one of the whitelisted roots.
Recommended location in ComfyUI\user instead of the node folder so it survives node updates.
Loader lookup order: env var → global ComfyUI locations → node folder.
Lines starting with # are treated as comments in the JSON file.
An allowed root permits saving in that folder and all subfolders; whitelist a deeper path to restrict more tightly.

Path and filename templates Placeholders supported in file_path and filename_prefix:

[time(%Y-%m-%d)] → formatted time (strftime)
[date] → YYYY-MM-DD
[datetime] → YYYY-MM-DD_HH-MM-SS
[unix] → epoch seconds
[guid] / [uuid] → random UUID4 hex
[model] → tries extra_pnginfo keys: model, checkpoint, ckpt_name, model_name; else unknown
[env(NAME)] → environment variable NAME

Examples

audio/[time(%Y-%m-%d)]
runs/[model]/[datetime]
D:/Exports/[env(USERNAME)]/[guid]

Installation

Install from ComfyUI Manager (search for Dehypnotic), or

Navigate to your ComfyUI custom_nodes directory:
```
cd ComfyUI/custom_nodes
```

Clone this repository:

git clone https://github.com/your-username/ComfyUI-Dehypnotic.git

Restart ComfyUI.

License

⚖️ Licensed under <a href="https://github.com/Dehypnotic/ComfyUI-Dehypnotic/blob/main/LICENSE">MIT</a>

Dehypnotic

Important: This is my only repository now, so delete any old folders (comfyui-aspect-ratio-advanced, comfyui-numbered-text, comfyui-range-to-string, or comfyui-dehypnotic-save-nodes) you may have in ComfyUI\custom_nodes to avoid conflicts.

ComfyUI-Dehypnotic Custom Nodes

Table of Contents

Nodes Overview

🧘 AspectRatio (Dehypnotic)

Key Features:

Inputs & Outputs:

🧘 Set Dehypnotic & Get Dehypnotic

Key Features:

How to Use:

🧘 NumberedText (Dehypnotic)

Key Features:

Inputs & Outputs:

🧘 Text (Dehypnotic)

Key Features:

Inputs & Outputs:

🧘 RangeToString (Dehypnotic)

Key Features:

Inputs & Outputs:

🧘 Save Audio (Dehypnotic)

Key Features:

Inputs & Outputs:

🧘 Save Images (Dehypnotic)

Key Features:

Inputs & Outputs:

🧘 FrameSave (Dehypnotic)

Key Features:

Inputs & Outputs:

🧘 Load Video (Dehypnotic)

Key Features:

Inputs & Outputs:

🧘 Save Video (Dehypnotic)

Key Features:

Key Inputs & Outputs:

🧘 SonicColor Noise (Dehypnotic)

Key Features:

Inputs & Outputs:

🧘 BrainwaveSync (Dehypnotic)

Key Features:

Inputs & Outputs:

Security and external save paths (ComfyUI Manager compliant)

Installation

License