Monta fluxos de IA generativa em nós para criar imagens, vídeos e áudio no PC ou na nuvem

por Comfy Org

Em portuguêsUsa inteligência artificial
Nota dos leitores
Sem avaliaçõesAvaliar
Downloads pelo Flipters
0
Estrelas no GitHub
148 mil

Última versão: v0.46.0, de 18 de novembro de 2026

Sobre o ComfyUI

ComfyUI é um programa de código aberto da Comfy Org para gerar imagens, vídeos, áudio e modelos 3D com IA, ligando cada etapa em uma interface de nós. Atende criadores e equipes técnicas que querem controle sobre modelos e parâmetros.

Roda no computador pelo app Comfy Desktop, pela versão portátil ou por instalação manual, e também no Comfy Cloud, serviço pago com GPUs na nuvem. Uma placa de vídeo dedicada é recomendada.

Principais recursos

  • Editor de workflows em nós para imagem, vídeo, 3D e áudio
  • Roda modelos abertos em GPUs a partir de 4 GB de VRAM
  • App Mode mostra workflows complexos em uma tela simples
  • Mais de 5.000 extensões criadas pela comunidade
  • Partner Nodes dão acesso a modelos fechados, como Nano Banana

Arquivos para baixar

Instaladores publicados pelo fabricante, separados por sistema. Na dúvida, use o botão principal: ele leva à página oficial de download.

Windows
ArquivoVersãoTamanhoBaixar
Windows (arm64) .7zComfyUI_windows_portable_nvidia_arm64.7zv0.46.01,8 GBBaixar Windows (arm64) .7z
Windows .7zComfyUI_windows_portable_amd.7zv0.46.01,6 GBBaixar Windows .7z
Windows .7zComfyUI_windows_portable_intel.7zv0.46.01,4 GBBaixar Windows .7z
Windows .7zComfyUI_windows_portable_nvidia.7zv0.46.01,9 GBBaixar Windows .7z
Windows .7zComfyUI_windows_portable_nvidia_cu126.7zv0.46.01,8 GBBaixar Windows .7z

Novidades

Novidades da versão v0.46.0Publicada em 18 de novembro de 2026. Notas do GitHub, em inglês.Mostrar

What's Changed

  • [Partner Nodes] feat(LTX): add AlphaGen, SDR to HDR, Reframe and Upscale nodes by @bigcat88 in https://github.com/Comfy-Org/ComfyUI/pull/17664
  • [Partner Nodes] feat(MiniMax): add MiniMax H3 Image node by @bigcat88 in https://github.com/Comfy-Org/ComfyUI/pull/17661
  • Update workflow templates to v0.11.106 by @comfyui-wiki in https://github.com/Comfy-Org/ComfyUI/pull/17668
  • Fix for some workflows not working properly. by @comfyanonymous in https://github.com/Comfy-Org/ComfyUI/pull/17675
  • Conform Concatenate Video clips to the first clip's frame rate and audio (FRONT-3775) by @jtydhr88 in https://github.com/Comfy-Org/ComfyUI/pull/17621
  • [Partner Nodes] feat(Wan): move Wan nodes to Comfy Router by @bigcat88 in https://github.com/Comfy-Org/ComfyUI/pull/17639
  • [Partner Nodes] feat(Krea): move Krea nodes to Comfy Router by @bigcat88 in https://github.com/Comfy-Org/ComfyUI/pull/17691
  • [Partner Nodes] chore(ByteDance): drop the Seedance 2.5 4k enterprise-only wording by @bigcat88 in https://github.com/Comfy-Org/ComfyUI/pull/17699
  • Bump comfyui-frontend-package to 1.59.1 by @comfy-pr-bot in https://github.com/Comfy-Org/ComfyUI/pull/17553
  • Remove directml support. by @comfyanonymous in https://github.com/Comfy-Org/ComfyUI/pull/17711
  • [Partner Nodes] feat(Gemini-LLM): add Hydrogen/Argon and Carbon models by @bigcat88 in https://github.com/Comfy-Org/ComfyUI/pull/17724
  • Update workflow templates to v0.11.107 by @comfyui-wiki in https://github.com/Comfy-Org/ComfyUI/pull/17725
  • Support saving single channel PNG in the save image advanced node. by @comfyanonymous in https://github.com/Comfy-Org/ComfyUI/pull/17727
  • Add new node to split image channels by @alexisrolland in https://github.com/Comfy-Org/ComfyUI/pull/13548
  • [Partner Nodes] feat(HeyGen): add 1080p, 30s clips and auto duration to HeyGen Video 1.0 nodes by @bigcat88 in https://github.com/Comfy-Org/ComfyUI/pull/17721
  • [Partner Nodes] fix(FishAudio): show Drama 3 full price in the badge by @bigcat88 in https://github.com/Comfy-Org/ComfyUI/pull/17718
  • [Partner Nodes] chore(OpenRouter): remove retired Qwen3.6 Max Preview model by @bigcat88 in https://github.com/Comfy-Org/ComfyUI/pull/17737
  • [Partner Nodes] feat(Luma): move Luma nodes to Comfy Router by @bigcat88 in https://github.com/Comfy-Org/ComfyUI/pull/17714
  • [Partner Nodes] feat(Anthropic): add Claude Fable 5.2 to the Claude node by @bigcat88 in https://github.com/Comfy-Org/ComfyUI/pull/17738
  • Fix some cases where apg affected output with cfg 1.0 by @comfyanonymous in https://github.com/Comfy-Org/ComfyUI/pull/17741
  • Show Kling Avatar task IDs on polling timeout by @wei-hai in https://github.com/Comfy-Org/ComfyUI/pull/17740
  • Show Topaz request IDs on polling timeout by @wei-hai in https://github.com/Comfy-Org/ComfyUI/pull/17742
  • Add a mesh input to Create Camera Info to frame any mesh by @kijai in https://github.com/Comfy-Org/ComfyUI/pull/17730
  • Support embedded tokenizers for vocabulary-pruned Qwen checkpoints by @rattus128 in https://github.com/Comfy-Org/ComfyUI/pull/17633
  • Video Deflicker node and Faster RAFT optical flow implementation by @kijai in https://github.com/Comfy-Org/ComfyUI/pull/17754
  • Don't hardcode video token id. by @comfyanonymous in https://github.com/Comfy-Org/ComfyUI/pull/17759
  • Add warning if someone is not using cuda 13 on a GPU that support it. by @comfyanonymous in https://github.com/Comfy-Org/ComfyUI/pull/17760
  • [Partner Nodes] feat(OpenRouter): add Meta Muse Wonder, Spark 1.3 and Ember models by @bigcat88 in https://github.com/Comfy-Org/ComfyUI/pull/17744
  • [Partner Nodes] feat(ByteDance): add Seedance 2.5 Flash to the Seedance 2.5 nodes by @bigcat88 in https://github.com/Comfy-Org/ComfyUI/pull/17764
  • Update workflow templates to v0.11.109 by @comfyui-wiki in https://github.com/Comfy-Org/ComfyUI/pull/17765
  • feat: Support PAN2 (CORE-719) by @kijai in https://github.com/Comfy-Org/ComfyUI/pull/17755
  • Make the cuda version warning
Ver a versão no GitHub

Avaliações de quem usa

Ainda sem avaliações

Usou o ComfyUI? Conte para quem está escolhendo.

Avalie o ComfyUI

Sua nota (obrigatória)

De 20 a 3.000 caracteres.0/3.000

Alternativas ao ComfyUI

  • Stable Diffusion web UI

    AUTOMATIC1111 · Imagem e design com IA

    Interface web local para gerar e editar imagens com modelos Stable Diffusion

    • 165,4 mil no GitHub
  • InvokeAI

    Comunidade InvokeAI · Imagem e design com IA

    Gera e edita imagens com IA no seu computador, com canvas em camadas e workflows em nós

    • 29,1 mil no GitHub
  • Fooocus

    lllyasviel · Imagem e design com IA

    Gera imagens com modelos SDXL no seu PC a partir de prompts simples, sem ajustes técnicos

    • 54 mil no GitHub
Ver todos em Imagem e design com IA

Fontes

Ficha escrita pela redação do Flipters a partir das fontes acima, revisada em 21 de novembro de 2026. Preços e versões mudam: confira no site oficial antes de comprar. Como trabalhamos.