comfyui-mcp/packs/z-image-turbo-inpainting/manifest.yaml

Local-first, agent-native control plane for ComfyUI — MCP server + autonomous sidebar agent that drives your live graph in natural language on ANY LLM: Claude/ChatGPT/Gemini on your subscription (no API key), free local models via Ollama (fully offline),

# Z-Image-Turbo inpainting install manifest — consumable by apply_manifest.
# Sliced from the z-image-turbo monolith's "Z-IMAGE TURBO INPAINTING" pipeline
# (un-bypassed standalone). No film-grain / sharpen post-proc, so no librosa pin.
# Mirror base: https://huggingface.co/Aitrepreneur/FLX/resolve/main  (third-party)
#
# Do NOT copy the monolith's custom_nodes list: this standalone graph only uses
# GGUF loaders, Power Lora Loader, Easy-Use cache/GPU cleanup, and MaskBlur+
# (#2484). The eight packs dropped here (KJNodes, UltimateSDUpscale, wlsh,
# vrgamedevgirl, RES4LYF, SeedVarianceEnhancer, Detail-Daemon, controlnet_aux)
# supply no node in workflow.json.

pip: []

custom_nodes:
  - https://github.com/city96/ComfyUI-GGUF
  - https://github.com/rgthree/rgthree-comfy
  - https://github.com/yolain/ComfyUI-Easy-Use
  - https://github.com/cubiq/ComfyUI_essentials

models:
  # UNet (GGUF) — Z-Image-Turbo DiT
  - url: https://huggingface.co/Aitrepreneur/FLX/resolve/main/z_image_turbo-Q8_0.gguf
    local_path: unet/z_image_turbo-Q8_0.gguf
  # Text encoder — Qwen3-4B (CLIP type qwen_image)
  - url: https://huggingface.co/Aitrepreneur/FLX/resolve/main/Qwen3-4B-UD-Q6_K_XL.gguf
    local_path: text_encoders/Qwen3-4B-UD-Q6_K_XL.gguf
  # VAE — z-image VAE (saved as z-image-ae.safetensors to avoid the Flux ae clash)
  - url: https://huggingface.co/Aitrepreneur/FLX/resolve/main/ae.safetensors
    local_path: vae/z-image-ae.safetensors