Krea 2 ComfyUI Workflow: Optimized for Low VRAM and Human Anatomy
Sarcastic-Tofu · reddit · 2026-07-12
A developer shared a ComfyUI text-to-image workflow based on the Krea 2 model. Compared to Flux.2 klein's tendency to break human anatomy and Z-Image Turbo's occasional detail loss, Krea 2 excels in visual coherence.
Key features of this workflow include:
- Low barrier and high consistency: Delivers great results using the Q2 quantized version on just 8GB VRAM. It natively supports blending up to 10 reference images, drastically improving character consistency.
- Uncensored and user-friendly: Generates NSFW content without needing special nodes. Built-in nodes automatically save generation parameters and metadata into a txt file.
- Model acquisition: Required models are available on HuggingFace, with companion LoRAs on CivitAI.
More from Multimodal
- Short film 'Odissey of Aeons' made with MiniMax H3 video generation — Musikalandia · 2026-10-27
- Remiqora Video: open-source local studio turns one idea into 4K chained AI videos — inikolax · 2026-10-27
- $5k kale-img-1 fine-tuning contest won with MOSS-Image base, SDXL trails far behind — ostrisai · 2026-10-27
- HN discussion: How far image generation has come since August 2022 — hunterbrooks · 2026-10-27
- FLUX.3 Edit nails a -90° car rotation on first try with just two reference images — EvilBot-666 · 2026-10-27
- Fan-Made AI Series 'Wolverine: Trail of Tears' Episode One Released — ArtificialLightPixel · 2026-10-27