ComfyUI_Primere_Nodes

ComfyUI_Primere_Nodes
★ 141

模型编排提示词管理多模型兼容网络集成
基于Primiere Model Control的系统性节点包，自动适配25+模型类型的采样、CFG、VAE、CLIP、LoRA等生成参数，集中管理提示与模型资源。
💡 按模型类型自动配置生成参数并集中管理提示与LoRA
🍴 19 Forks💻 Python🔄 2026-03-26
📦
网盘下载
复制链接后前往夸克网盘下载
https://pan.quark.cn/s/2df45d172dc1
📦 requirements.txt
deepface
Pillow
PyYAML
Requests
addict
aiohttp
chardet
dynamicprompts
ftfy
ipdb
numpy
opencv_contrib_python
opencv_python
opencv_python_headless
pandas
pefile
piexif
pipreqs
pycocotools
pyexiv2
pytorch_lightning
pytz
regex
segment_anything
setuptools
supervision
timm
tomli
torch
torchvision
tqdm
transformers
ultralytics
yapf
openai-clip
huggingface_hub
gguf>=0.9.1
imagesize
sentencepiece
accelerate
cpm-kernels
diffusers
bitsandbytes
peft>=0.10.0
pyrallis
mmengine
open_clip_torch
omegaconf
fal-client
anthropic
openai
google-genai>=1.65.0
dotenv
elevenlabs
python-magic-bin
📄 README

Primere nodes for ComfyUI

Git link: https://github.com/CosmicLaca/ComfyUI_Primere_Nodes

Primere Youtube channel: https://www.youtube.com/@PrimereComfydev/videos

Install required 3rd party nodepack dependency: https://github.com/city96/ComfyUI_ExtraModels

Nodepack under development. Manual written by AI, please open issue if something wrong or missing. If you upgrade, check attached new workflows or use git to downgrade to previous version if something failed.

ComfyUI core functions requirement: This nodepack uses new ComfyUI core functions. Update your ComfyUI installation before using. Use git pull on your ComfyUI folder.

All example workflows in the Workflow folder of the node root.

  • Within Workflow folder you will find subfolders:
  • NewFE: workflows compatible with current ComfyUI front-end
  • Development: workflows for node testing only

    • About This Nodepack

    Primere Nodes is not a collection of individual utility nodes. It’s a consistent full-system solution architected around a central orchestrator (Primiere Model Control) that automatically adapts all generation parameters (sampling, CFG, VAE, CLIP, encoders, LoRAs, refiners) based on selected model type or individual checkpoint.

    Why the complexity? Supporting 25+ diverse model architectures (SD1, SDXL, Flux, Hunyuan, etc.) with conflicting requirements demands a system-level approach. Each model expects different schedulers, sigma ranges, CFG scales, CLIP handling, and attention mechanics. Manual parameter tweaking per model is unsustainable at scale. Primere solves this by:

  • Saving per-model settings to JSON configs
  • Auto-loading correct settings when checkpoint selected
  • Enforcing compatibility between model type and network types (removing incompatible LoRAs, etc.)
  • Providing override capability for testing/optimization, then re-saving for future use

    • The workflow appears complex because it handles real-world generation scenarios (model switching, quality filtering, preview building, multi-stage refinement) that naive single-sampler approaches ignore.

    [!TIP]

    ## Something different: Universal API node 🚀

    Explore the Uniapi quick manual: PrimereApiProcessor (Uniapi) guide.

    Minimal workflow

    Detailed manual for included nodes

    Core generation pipeline. Single prompt input with full model concept support.

    Central Orchestrator: Primiere Model Control node reads selected checkpoint, auto-loads saved settings (sampler, CFG, steps, VAE, CLIP, encoders, LoRAs, refiners), and distributes control_data tuple to all downstream nodes for automatic parameter application.

    Supported model concepts: SD1, SD2, SDXL, Illustrious, SD3, StableCascade, Chroma, Z-Image, Turbo, Flux, Nunchaku, QwenGen, QwenEdit, WanImg, KwaiKolors, Hunyuan, Playground, Pony, LCM, Lightning, Hyper, PixartSigma, SANA1024, SANA512, AuraFlow.

    Future support: HiDream, Mochi, WanT2V, WanI2V, Cosmos, Flux2, SSD, SegmindVega, KOALA, StableZero, SV3D, SD09, StableAudio, LTXV.

    Minimal workflow features:

  • Central model orchestrator: Auto-configures sampler, CFG, steps, VAE, CLIP, encoders, attention, LCM, speed or SRPO LoRAs or refiners per model type or model concept type
  • Automatic model keyword insertion to prompt
  • Prompt selector to any prompt sources
  • Prompt can be saved to CSV file directly from the prompt input nodes
  • CSV and TOML file readers for saved prompts, automatically organized, saved prompt selection by preview image (if preview created)
  • Randomized latent noise for variations
  • Prompt encoder with selectable custom clip model, long-clip mode with custom models, advanced encoding, injectable internal styles, last-layer options
  • Sampler with variation extender and Align Your Steps features
  • A1111 style network injection supported by text prompt (LoRA, LyCoris, Hypernetwork, Embedding)
  • Network tag cleaner: Auto-removes incompatible network tags per model architecture
  • Automatized and manual image saver with optional preview saver for checkpoint selectors and saved .csv prompts
  • Aesthetic scorer for final image quality assessment
  • Upscaler (selectable Ultimate SD and hiresFix)
  • Dynamic prompt support
  • Two-stage refiner support (quality refinement stage)
  • Speed LoRAs: LCM, Lightning, Hyper for fast generation

    • Basic workflow

    Detailed manual for included nodes

    Professional prompt development workflow. Extended prompt management for testing and iteration. Adds postprocessing and style injection capabilities to Minimal workflow.

    Same central orchestration and model support as Minimal workflow.

    Basic workflow features:

    Same as Minimal workflow plus:

  • 12 prompt inputs with 1-click selector for prompt switching
  • Style injection node: Select result art style for prompt injection
  • Photoshop-style postprocessing nodes: Image manipulation and effects
  • Histogram for result checking
  • Efficient workflow for prompt developers testing multiple variations with integrated finishing

    • Basic Production workflow

    Detailed manual for included nodes

    Full production pipeline with styling, refinement, and selective output.

    Same central orchestration and model support as Minimal workflow.

    Basic Production workflow features:

    Same as Basic workflow plus:

  • 19 prompt sources for special cases (Daily Challenges, Articles, etc.)
  • Style block: Style Pile, Midjourney, Emotions, Camera Lens (prompt injection styling)
  • 4 separated refiner detailer blocks: Face, Eye, Mouth, Hands
  • Selective image saver: saves only images exceeding user-defined aesthetic score threshold (portfolio filtering)

    • Workflow Comparison

    | Feature | Minimal | Basic | Basic Production |

    |——————————–|—————–|——-|——————|

    | Primary Use | Core generation | Prompt dev + postprocessing | Production pipeline |

    | Prompt Inputs | 4 | 12 | 19 |

    | Prompt Selector | 1-click | 1-click | 1-click |

    | Model Concepts | 25+ | 25+ | 25+ |

    | Central Model Orchestrator | ✓ | ✓ | ✓ |

    | Model Keyword Insertion | ✓ | ✓ | ✓ |

    | CSV/TOML Readers | ✓ | ✓ | ✓ |

    | Network Injection | ✓ | ✓ | ✓ |

    | Network Tag Cleaner | ✓ | ✓ | ✓ |

    | Variation Extender | ✓ | ✓ | ✓ |

    | Image Saver | ✓ | ✓ | ✓ |

    | Aesthetic Scorer | ✓ | ✓ | ✓ |

    | Speed LoRAs | ✓ | ✓ | ✓ |

    | Two-Stage Refiner | ✓ | ✓ | ✓ |

    | Style Injection Node | — | ✓ | ✓ |

    | Postprocessing Nodes | — | ✓ | ✓ |

    | Histogram | — | ✓ | ✓ |

    | Style Block | — | — | ✓ |

    | Refiner Detailers | — | — | ✓ (Face, Eye, Mouth, Hands) |

    | Selective Saver | — | — | ✓ |