Preview for FantasyTalking on ComfyUI
FantasyTalking on ComfyUI workflow diagram

Run this workflow on InstaSD

Get started in minutes! Run this ComfyUI workflow online - no setup required.

Description

Bring Your Portraits to Life with FantasyTalking on ComfyUI

FantasyTalking by Alibaba makes it easy to transform static portraits into lively talking characters. By syncing facial expressions and lip movements with audio, this workflow creates realistic animations that make characters feel truly alive. Whether you're working with realistic avatars or fun cartoon characters, FantasyTalking delivers smooth, high-quality talking videos that are perfect for a variety of creative projects.

🎯 Features

  • Realistic Lip Sync – Get natural mouth movements that perfectly match the audio.
  • Smooth Motion – Animate facial expressions and body movements for more dynamic results.
  • Works with Any Style – From realistic faces to cartoon avatars, you can animate any character.
  • High-Quality Video – Generate full-body, half-body, or close-up videos that look and feel lifelike.

💡 Use Cases

  • Character Animation – Bring virtual influencers, game characters, or avatars to life with talking animations.
  • AI Research – Dive into the world of audio-visual generation and experiment with lip-sync techniques.
  • Video Content – Create engaging talking videos for social media, marketing, or entertainment.

⚙️ How It Works

  1. Launch the Workflow – Open the pre-configured ComfyUI instance with FantasyTalking.
  2. Upload Your Audio & Image – Add your desired audio and a portrait of the character.
  3. Generate & Preview – Let InstaSD’s cloud infrastructure work its magic to create the video.
  4. Download Your Video – Save your finished video for use in projects, social media, or content creation.

With FantasyTalking, transforming a simple portrait into a talking character is easier than ever. Whether it’s for a fun project, an AI experiment, or just to make your characters come to life, get started with FantasyTalking today!

Models

FileDestinationSource
fantasytalking_fp16.safetensors/ComfyUI/models/diffusion_modelsDownload
Wan2_1_VAE_bf16.safetensors/ComfyUI/models/vaeDownload
umt5-xxl-enc-bf16.safetensors/ComfyUI/models/text_encodersDownload
clip_vision_h.safetensors/ComfyUI/models/clip_visionDownload
Wan2_1-I2V-14B-720P_fp8_e4m3fn.safetensors/ComfyUI/models/diffusion_modelsDownload

Nodes

NoteWanVideoVRAMManagementCLIPVisionLoaderWanVideoDecodeWanVideoClipVisionEncodeWanVideoSamplerMarkdownNoteLoadWanVideoT5TextEncoderPreviewImageSetNodeGetNodeWanVideoImageToVideoEncodeINTConstantWanVideoBlockSwapDownloadAndLoadWav2VecModelFantasyTalkingWav2VecEmbedsVHS_VideoCombineWanVideoTeaCacheCreateCFGScheduleFloatListLoadImageWanVideoModelLoaderWanVideoTorchCompileSettingsWanVideoVAELoaderFantasyTalkingModelLoaderLoadAudioWanVideoTextEncodeImageResizeKJv2