MOSS-Image ecosystem roundup: Moebius LoRA, 4GB VRAM build and new ComfyUI workflows
optimisticalish · reddit · 2026-10-31
A Reddit roundup of the latest MOSS-Image ecosystem resources:
- Moebius-style LoRA: no trigger word needed; best at 1.2 strength, 1552px widescreen, Spectrum acceleration with 15 steps — more steps confuse the output. Includes a full ligne claire prompt template and tips like outlining distant clouds and desert dust with small dots.
- MOSS-Image-SFTclear fine-tune: reduces noise for a cleaner look and vivid colors, aimed at anime illustrations; 5.8GB INT8 checkpoint.
- Unmodel slider LoRA (15GB) to nudge faces away from the 'doll-face/glamour' look.
- ComfyUI-MOSS-Card-Layout: three/four-panel character cards with a searchable 100-outfit wardrobe, 60-face library, live thumbnails and a persistent personal library.
- A working ComfyUI workflow for AMD RX 6900 XT 16GB, plus a 4-bit 5B DiT build (no text encoder) that runs on 4GB cards.
- A MOSS-Image Prompt Enhancer node shipping with a .MD skill.
More from coding & agent
- Playwright MCP Server lets AI assistants drive remote browsers — modelcontextprotocol · 2026-10-31
- zero-core-tools: MCP connector bundling scraping, code review, sentiment — modelcontextprotocol · 2026-10-31
- Codex power user shows 71.5B lifetime tokens and a 262-day daily streak — TheMoonMidas · 2026-10-31
- ComfyUI-Dials: prompt pairs become 14KB sliders controlling video motion & mood, no LoRA training — MatlowAI · 2026-10-31
- Dev Asks GPT-6 Astra to Motorize an Office Chair, Then Builds It in Real Life — sdand · 2026-10-31
- Truo: read-only MCP server pipes live Australian bank data into AI assistants — truo_au · 2026-10-31