Oasis Suite v1.5: Video Oasis Viewer and LTX2.3 Oasis for ComfyUI

Oasis Suite v1.5 is a custom node suite for ComfyUI by NikoDemon80, built on the original Image Oasis. It expands into three all-in-one nodes: Video Oasis Viewer (preview-first Save Video with in-node player and scene bar) and LTX2.3 Oasis (LTX 2.3 video generation with Prompt Beats, LoRA stack, and audio modes).

What's New in v1.5

v1.5 transforms Image Oasis into a compact suite of three nodes that share the same player language, save habits, and UI patterns:

Node	Class ID	Status	Purpose
Image Oasis	`ImageOasis`	v1.4 carried forward	Multi-architecture image generation with LoRA stack, enhancer, refiner, upscale
Video Oasis Viewer	`VideoOasisPreview`	New	Preview-first Save Video with in-node player, scene bar, clip, and movie creation
LTX2.3 Oasis	`LTX23Oasis`	New	All-in-one LTX 2.3 T2V/I2V with Prompt Beats, audio, LoRA stack, and GGUF support

Video Oasis Viewer

A preview-first Save Video replacement. Connect a VIDEO input and the node encodes to temp and plays it in-node — nothing hits your output folder until you press Save.

Key features:

In-node player with scrub bar, frame-step, mute, speed control, and lightbox (scroll zoom, drag pan)
Frame drag — pause on any frame and drag it onto any image input on the graph (Load Image, refs, LTX guides)
Scene bar — up to 24 clips, click to recall, delete, long-press reorder. Load from output/ via the + button
Clip — mark in [ and out ], then Clip to create a trimmed copy
Create Movie — concat every saved bar clip into a single file (stream-copy when possible)
Encode/Save with format, codec, quality, and save prefix settings

LTX2.3 Oasis

All-in-one LTX 2.3 video generation in the Image Oasis UI shape:

Text-to-Video and Image-to-Video modes
↻ Continue from viewed video — chain clips by continuing from the last frame of whatever is in the viewer
Prompt Enhancer — resident LLM (GGUF) for expanding prompts
LoRA Stack — add, reorder, toggle, and set trigger words
Prompt Beats — multi-prompt timeline with optional guide images at each beat
Audio modes: Off / Generate / File (audio-driven video)
GGUF diffusion support via optional ComfyUI-GGUF package
Scene bar persistence across ComfyUI tab switches and reloads

Image Oasis (Carried Forward)

The original all-in-one image generation node continues with the same capabilities: architecture registry (FLUX, SD, Krea, etc.), prompt enhancer, LoRA stack, refiner, upscale, history strip, bypass, and CivitAI-hash support.

Installation

Clone or update the repository:

cd ComfyUI/custom_nodes
git clone https://github.com/NikoDemon80/ComfyUI-Image-Oasis.git
# Or pull latest for existing installations
cd ComfyUI-Image-Oasis && git pull

Then install Python dependencies:

pip install -r requirements.txt

Restart ComfyUI and hard-refresh the browser (Ctrl+F5).

Optional Dependencies

llama-cpp-python (CUDA/Metal) for the GGUF prompt enhancer
ComfyUI-GGUF for GGUF diffusion models
ComfyUI-KJNodes for LTX2 NAG when CFG is 1 and a negative prompt is set

Why It Matters

The Oasis Suite fills a gap in the ComfyUI ecosystem by providing a cohesive, all-in-one experience for both image and video generation. The Video Oasis Viewer's in-node preview and scene bar workflow — combined with LTX2.3 Oasis's Prompt Beats and audio-driven video — makes it a compelling toolkit for AI video creators who want a streamlined interface without leaving ComfyUI.

The suite is released under GPL-3.0-or-later.