artokun/comfyui-mcp

AI-Toolkit LoRA Trainer trains custom LoRAs with ostris AI-Toolkit for WAN 2.2/2.1 and Z-Image. Use it to prepare datasets, tune training settings, and generate a LoRA for use in ComfyUI.

Anima Base generates anime and illustration images with the ANIMA 1.0 model in ComfyUI. Use it for Danbooru tags or natural-language prompts, low-VRAM generation, and anime inpainting with Anima-LLLite ControlNet.

Citron Anima LoRA Trainer is a local Gradio UI for training LoRA adapters on the Anima base model with kohya sd-scripts. Use it to prepare datasets, tune training settings, and export a LoRA for the anima-base workflow.

Civitai discovers Civitai models with comfyui-mcp search, downloads them into ComfyUI, and uses trigger words for generation. It is used to find checkpoints, LoRAs, or embeddings and install the right version locally.

Color Correction measures image or video frames with analyze_color so you can diagnose washout, clipping, saturation, and color cast from stats instead of guessing. Use it when a render looks flat or off, and re-measure after contrast, levels, or match fixes.

ComfyUI Core explains ComfyUI workflow JSON, node types, and common pipeline patterns. Use it to understand, edit, validate, and save workflows with MCP tools.

ComfyUI Frontend Extensions helps you author ComfyUI v2 web UI extensions with @comfyorg/extension-api, including nodes, widgets, sidebar panels, commands, and hotkeys. Use it when writing or migrating frontend extension code for ComfyUI.

ComfyUI Launch Flags helps choose startup flags for VRAM use, caching, and attention backends. Use it when a graph OOMs, shared VRAM hurts performance, models need to stay resident, or a model needs a specific attention mode.

ComfyUI Node Pack Skill Generator creates a Claude Code skill for a ComfyUI custom node pack from a registry ID or GitHub URL. Use it when you want a generated SKILL.md with node specs, examples, and source notes.

ComfyUI Node Registry covers writing ComfyUI custom nodes and publishing them to the Comfy Registry. Use it when you need the node package structure, pyproject.toml metadata, comfy-cli publishing, or CI setup.

Debug Render localizes wrong but completed ComfyUI outputs by rendering one branch to an output node and inspecting intermediate images, latents, masks, and preprocess maps. Use it when a final image or video has artifacts, wrong composition, or a conditioner is not taking effect.

Director orchestrates a story-to-video production pipeline from scene planning through frames, video clips, and final ffmpeg concatenation. Use it when you need consistent character-driven short film generation in ComfyUI.

Ernie Image builds Baidu ERNIE-Image and ERNIE-Image-Turbo text-to-image workflows for ComfyUI. Use it for multilingual text rendering, posters, manga layouts, and complex multi-object prompts; it also supports denoise-based image refinement.

Flux Text-To-Image Workflows builds ComfyUI setups for Flux image generation, including Flux.1 Dev SRPO, Flux 2 Klein 9B, Turbo LoRAs, and FluxGuidance. Use it when you need the right loaders, conditioning, and sampler settings for Flux models.

Ideogram Ultra builds local Ideogram 4 txt2img and img2img workflows in ComfyUI. Use it for structured JSON prompts, text rendering, and layout control with dual-model guidance.

Installer Packs helps you install a ComfyUI model family from a manifest-driven pack, or derive a new pack from an upstream installer or workflow JSON. It is used when building, validating, and sharing packs back upstream.

Krea 2 Identity Edit swaps or transfers clothing onto a person while preserving their face, pose, and background. Use it for outfit transfer, virtual try-on, and dressing one person in garments from another image.

Krea 2 Text To Image Workflows builds ComfyUI txt2img packs for Krea 2 Turbo, including manual prompts, JSON area prompting, and a two-pass detail-boost workflow. It is used when generating images with native krea2 encoders, Qwen VAE, and 8-step turbo settings.

LTX Director Timeline edits the hidden timeline_data JSON for the LTXDirector node. Use it to add, move, retime, or remove image, text, audio, and motion segments when the node's DOM buttons can't be clicked.

LTX V2 Video Workflows builds Lightricks LTX-2 and LTX-2.3 ComfyUI graphs for text-to-video, image-to-video, audio sync, and two-stage upscaling. Use it when setting up bundled or GGUF model paths and quality-tuned workflows.

LoRA Manager authors ComfyUI LoRA Manager nodes from the panel and explains which ones panel_add_node can create. Use it when adding loaders, stackers, prompts, or text nodes, or when autocomplete text nodes are refused.

Local LLM Free runs the ComfyUI agent on Ollama with fine-tuned local models, so you can use it offline without an API key or subscription. Use it when you want free local inference or need help choosing a model and VRAM size.

MiniMax H3 Video builds local ComfyUI workflows for MiniMax H3 text-to-video, image-to-video, and reference-to-video. Use it when you want to download the INT8 models, use turbo LoRAs for low VRAM, or follow the official prompting guide by link.

Model Compatibility Matrix maps loaders, resolutions, samplers, CFG, VAE, ControlNet, and LoRA support across SD 1.5, SDXL, Flux, SD3, and video models. Use it when choosing settings or checking whether assets work together.

Model Registry lists curated download URLs and target folders for ComfyUI models such as checkpoints, VAEs, text encoders, and LoRAs. Use it when a workflow reports a missing model or when setting up a new machine.