# A quick Minimax H3 news round-up - 1st November 2026 — retained record

> Retained reddit record from reddit:comfyui.

Canonical: https://highsignal.app/data/records/204b5642715f0eda

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:comfyui)  published 2026-11-01 15:09 ingested 2026-11-02 02:41

[original source ↗](https://www.reddit.com/r/comfyui/comments/1xl5evm/a_quick_minimax_h3_news_roundup_1st_november_2026/)

## retained text



```
After a long pause, another quick Minimax H3 news and goodies round-up, for those who may have missed some items. 

-> A Technicolor Dreams LoRA. "The aim is to give dialogue, music, ambience, and imagery a shared cinematic character [... namely, that of U.S and British] "1930s–1950s cinema".

https://huggingface.co/Granddyser/minimax-h3-technicolor-dreams

-> MiniMax-H3-LoRA-LineartAnime. Line-art colorisation using H3, aided by DiffSynth-Studio.

https://huggingface.co/hakandinar/MiniMax-H3-LoRA-LineartAnime

-> H3 Recast... "swaps up to three people in a clip in one pass. The motion, the scene and the clip's own audio stay." See also RecastRoom Studio for a simpler 'same take, new lead actor' type of re-casting.

https://github.com/Bambushu/h3-recast

https://github.com/jmanhype/recastroom-studio

-> 35 minutes with DK Berry on YouTube, discussing the finer points of correct prompting for Minimax H3.

https://www.youtube.com/watch?v=T_ZRyr1VNZs

-> Going up? H3 ComfyUI nodes for uppainting. Similar to inpainting and outpainting, but it goes... up. Which I guess has its uses for mobile phone screens, where perhaps you want video content at the bottom plus a text or UI overlay at the top.

https://github.com/yangyyyy15/ComfyUI-H3UpperOutpaint

-> Is your flapple widget not wiggling into the woggle socket? Minimax-H3-VBVR aims to improve... "object consistency and character action logic, reducing unintended object duplication, clipping, and interpenetration during motion and interactions".

https://huggingface.co/LiconStudio/Minimax-H3-VBVR

-> MotionBridge for ComfyUI. Input a motion video, get a 3D figure motion file (SOMA-77 joints, aka *somaskel77* format), which is then cleaned and can be retargeted to apparently fit "any humanoid rig" in 3D figure software.

https://github.com/powerhouse90/motionbridge

-> Stage. "Build cinematic camera movements around a neutral mannequin in a 3D scene, then export them as a guide video (.MP4) plus an English text prompt to drive the camera" in Minimax H3.

https://github.com/aledige63/Stage

-> ComfyUI-Minimax-H3-Komikku. Run a video, have a workflow turn it into a basic comic-book page, then tweak the page with a slick node UI with in ComfyUI.  Tested at a basic level, and working.

https://github.com/quinteroac/ComfyUI-Minimax-H3-Komikku

-> Another character-sheet auto-maker for H3. Input a still image, get a turnaround character sheet.

https://civitai.com/models/3022612/minimax-h3-character-sheet-or-turnaround-from-one-photo

-> ComfyUI-Picky, renders ten 'takes', then you choose the best. Appears to require a 24Gb VRAM card.

https://github.com/chenthedesignmaker/ComfyUI-Picky

-> Minimax-H3-Transparent-Video. Generate clean video cut-outs with an alpha channel, suitable for use in high-end video compositing software such as After Effects. "One generation, one node, no two-pass tricks".

https://github.com/SatoDive/Minimax-H3-Transparent-Video

-> Lumit, a free open-source Adobe After Effects clone.

https://lumitlab.com/ 

-> The excellent Openshot desktop open-source video editing software has updated. Version 4.1 allows a LLM such as Claude Code or Codex to auto-edit your video. Users can also type text directly onto the video and also change the font. The latter feature does not replace the existing easy titles/credits-making options in OpenShot.

https://www.openshot.org/blog/2026/10/25/openshot-41-direct-text-editing-auto-optimize-command-line-editing/

-> A workflow that claims to make H3 into a single-image Edit model. 2Mpx or higher, no trigger. Appears to be about re-posing a character rather than re-styling the image.

https://civitai.com/models/3028251/minimaxh3-single-frame-edit

-> And finally... yes, it is possible to use Openpose reference images with Minimax H3 Edit, for re-posing. You just need to align the prompt properly. See my demo workflow screenshot below... 

https://old.reddit.com/r/StableDiffusion/comments/1xknlam/images_with_references/19ryqsk4hvs0z/

~ OLD POSTS ~

https://jurn.link/dazposer/index.php/h3/ (Organised directory page, made from these H3 posts)

https://old.reddit.com/r/comfyui/comments/1x899wr/a_quick_minimax_h3_news_roundup_17th_october_2026/

https://old.reddit.com/r/comfyui/comments/1x5puo0/a_quick_minimax_h3_news_roundup_14th_october_2026/

https://old.reddit.com/r/comfyui/comments/1x3bouf/a_quick_minimax_h3_news_roundup_11th_october_2026/

https://old.reddit.com/r/comfyui/comments/1x1o42k/a_quick_minimax_h3_news_roundup_8th_october_2026/

https://old.reddit.com/r/comfyui/comments/1x0rdab/a_quick_minimax_h3_news_roundup_8th_october_2026/

https://old.reddit.com/r/comfyui/comments/1x02vw6/a_quick_minimax_h3_news_roundup_7th_october_2026/ (See 7th October post, for links to older posts)

https://old.reddit.com/r/comfyui/comments/1wqwpah/a_quick_minimax_h3_news_roundup_26th_september/ (See 26th September post, for links to even older posts)

https://old.reddit.com/r/comfyui/comments/1wgc4lj/a_quick_minimax_h3_news_roundup_15th_september/ (See 15th September post, for links to very old posts)

https://old.reddit.com/r/comfyui/comments/1w5i9iq/a_quick_minimax_h3_news_roundup_2nd_september_2026/ (See 2nd September post, for links to the starting posts)
```



document fetched 2026-11-02 00:36 · canonical https://reddit.com/r/comfyui/comments/1xl5evm/a_quick_minimax_h3_news_roundup_1st_november_2026
