NEURALDRIFT
HUB
Access 50 high-fidelity ComfyUI frameworks engineered for consistency, speed, and maximum creative control.
FEATURED ARCHITECTURES
Flux.1 Dev — Text to Image
The gold-standard text-to-image pipeline. Separate UNETLoader architecture. Best open-source image quality as of 2026.
Flux.1 Schnell — 4-Step Fast
4-step distilled Flux. Full quality in under 6 seconds. Use for rapid prompt iteration.
LTX Video 2.3 — Basic T2V
LTX Video 2.3 text-to-video. 25 frames, 768×512, 20 steps. Entry-level video generation.
Analysis Engine
Flux.1 Dev — Text to Image
The gold-standard text-to-image pipeline. Separate UNETLoader architecture. Best open-source image quality as of 2026.
Flux.1 Schnell — 4-Step Fast
4-step distilled Flux. Full quality in under 6 seconds. Use for rapid prompt iteration.
SDXL Base + Refiner
Classic SDXL two-pass pipeline. Base generates the latent, Refiner adds fine detail at 1024px.
Stable Diffusion 1.5 Default
The original ComfyUI default workflow. 4GB VRAM. Every node teaches a core concept.
HiDream — 17B Text to Image
HiDream-I1 17B parameter model. State-of-the-art image quality in 2025. 4 text encoders.
Hunyuan Image 2.1
Tencent Hunyuan DiT 2.1. Exceptional bilingual (Chinese + English) prompt support.
Lumina Image 2.0
Shanghai AI Lab flow-matching model. Fast convergence, strong aesthetic quality. 10GB VRAM.
Qwen Image Edit (Multimodal)
Qwen 2.5 VL 7B multimodal editor. Natural language image instructions. Understands scene content.
Chroma — Stylized No-Negative
Chroma model variant. No negative prompt needed. Strong for illustration and creative content.
AuraFlow Text to Image
AuraFlow v0.3. Fast open-source flow-matching. Competitive quality in under 30 steps. 8GB.
LTX Video 2.3 — Basic T2V
LTX Video 2.3 text-to-video. 25 frames, 768×512, 20 steps. Entry-level video generation.
LTX Video — Cinematic 9:16
97-frame vertical video at 768×1344 for YouTube Shorts and TikTok. Motion scale 1.2.
LTX Video — Action Chase
High motion chase sequence. 97 frames, motion scale 1.4 for maximum kinetic energy.
LTX Video — Fast Draft
15-step, 25-frame draft generation for prompt testing before committing to full renders.
AnimateDiff Simple Loop
SD 1.5 + AnimateDiff v3 motion module. 16 frames, 512×512 seamless loop. 8GB compatible.
AnimateDiff Seamless Loop
Infinite seamless loop optimized for social media. Glowing particle preset, 16 frames.
AnimateDiff Landscape Timelapse
Cinematic timelapse animation. 24 frames, cloud movement pre-prompted.
AnimateDiff 360 Product Spin
Smooth 360-degree product rotation for e-commerce and marketing content.
AnimateDiff Slow Zoom
Cinematic slow zoom into a scene. 16 frames, wide landscape format, dramatic fog.
Wan 2.1 Video — 480p T2V
Alibaba Wan 2.1 480p text-to-video. State-of-the-art transformer video generation.
4x ESRGAN Upscale
RealESRGAN 4x upscale for photorealistic images. Runs on 6GB. Load any image, get 4x output.
4x Anime Upscale
RealESRGAN x4 Anime optimized upscale. Preserves linework and flat colors.
SDXL Image-to-Image
Transform existing images with SDXL. Denoise 0.6 for strong style transfer while preserving composition.
SD 1.5 Style Transfer
Aggressive style transfer at denoise 0.75. Cyberpunk preset. Works on 6GB cards.
SDXL Sketch to Photo
Turn rough sketches into photorealistic images at denoise 0.85. High fidelity conversion.
SDXL Inpainting
Fill, replace, or repair image regions using mask input. Seamless blending at 1024px.
SD 1.5 Object Removal
Remove objects from images with SD 1.5 inpainting. Clean background fill at 512px.
SDXL Product Photography
Commercial product photography on white background. Studio lighting tuned for clean shots.
SDXL Architectural Visualization
Professional architectural exterior rendering. Daylight, photorealistic, ultra-detailed.
FLUX Portrait v2 — Editorial
High-contrast editorial portrait workflow using FLUX fp8. Vogue-style lighting.
ControlNet Canny — SDXL
Canny edge detection guides SDXL generation. Strength 1.0 for maximum structural adherence.
ControlNet Depth — SDXL
Depth map guides spatial composition. Strength 0.8 balances depth with creative freedom.
ControlNet OpenPose
Pose-guided generation from skeleton input. Lock human poses precisely across generations.
ControlNet Lineart — Anime
Anime lineart colorization. Line drawing input → fully colored anime illustration.
ControlNet Tile Upscale
ControlNet Tile for detail-preserving upscaling. Adds fine texture while respecting original.
SDXL Batch ×4 Variations
Generate 4 variations simultaneously. Concept art prompts pre-loaded. Same seed, 4 outputs.
SDXL Batch ×8 Grid
8 simultaneous variations at 768×768. Product design and concept iteration at scale.
SDXL Logo & Brand Design
Minimal vector-style logo generation. White background, professional brand identity.
SDXL Concept Art Sheet
Game character concept art with multiple view prompts. ArtStation quality preset.
FLUX Realistic Person
Photorealistic human portraits with natural skin texture. DSLR-quality output.
SDXL Interior Design
Scandinavian interior design visualization. Natural light, architectural photography style.
SD 1.5 Pixel Art Generator
16-bit RPG sprite generation. Clean pixel art on transparent backgrounds for game assets.
ACE Step 1.5 — Text to Song
Full AI song generation with lyrics, melody, and instrumentation. Text prompt to complete audio track.
FLUX Food Photography
Restaurant-quality food photography. Shallow DOF, warm lighting, appetizing styling.
SDXL Sci-Fi Environment
Cinematic sci-fi interior scenes. Space station, holographic displays, crew. Movie quality.
SDXL Abstract Art Generator
Flowing geometric abstract art. Gallery-quality output for prints and digital art.
SDXL Surrealist Dreamscapes
Salvador Dali-style surrealism. High symbolic detail, dreamlike compositions, smooth textures.
SDXL Fantasy Creatures
High-detail fantasy creature design. Dragons, griffins, and chimera concepts at 1024px.
SDXL Cyberpunk City
Bladerunner-style futuristic cities. Aerial night shots with neon glow and rain.
FLUX Cinematic Portrait v2
Film grain analog portrait workflow. Side lighting, moody atmosphere, cinematic feel.
UNLOCK PREMIUM
ARCHITECTURES
Join our elite node-engineering list. Receive 3 exclusive workflows and a 15% discount on the upcoming NeuralDrift Pro Training Program.