HEISS UI v0.17: free ComfyUI frontend adds inpainting, grouping and native upscale
LeanderDeBuhr · reddit · 2026-10-10
A solo developer released HEISS UI, a free image-first frontend for ComfyUI, with v0.17 shipping today.
Highlights:
- Video gen, reference images, and inpainting work with almost any model with zero setup; custom workflows also supported
- Auto-grouping stacks images from similar prompts into photo stacks
- Native within-generation upscale
- Auto-installs missing components (VAE, text encoders, nodes)
- Phone access over LAN with a custom mobile UI, plus favorites, search, prompt history, and LoRA stacks
Setup takes about 2 minutes; the author claims it's the best way to use Comfy 80% of the time.
More from Multimodal
- MiniMax-H3 hailed as favorite open release of the year, runs fully local — AdinaYakup · 2026-10-23
- Michal Malewicz experiments with depth maps in his creative workflow — michalmalewicz · 2026-10-23
- One-Shot Tunable Style Transfer on Flux3 Via Reference Latent + Node — Winter_unmuted · 2026-10-23
- PhysVidBench accepted at EMNLP 2026: do video gen models grasp physical commonsense? — aykuterdemml · 2026-10-23
- EMNLP 2026 paper teaches multimodal LLMs to get cartoons via incongruity-resolution supervision — aykuterdemml · 2026-10-23
- A gorgeous Midjourney style: sref code 820221253 for dreamy lavender watercolor looks — umesh_ai · 2026-10-23