Alibaba's Wan2.7 video generation model is now available directly in ComfyUI through the Partner Nodes system, delivering a significant upgrade to video creation workflows inside the popular open-source node editor. The integration, announced on April 3, 2026, requires ComfyUI v0.18.5 and covers image-to-video, text-to-video, video editing, and reference-based generation.

What Happened

The ComfyUI team released native support for Wan2.7 as a Partner Node integration. Wan2.7 is a comprehensive upgrade over the previous v2.6 release from Alibaba's Wan video project, with improvements spanning image quality, audio handling, motion dynamics, stylization, and temporal consistency.

The supported capabilities are extensive:

  • Image-to-Video: First-frame driven, first-plus-last-frame driven, and audio-driven image animation
  • Text-to-Video: Prompt-based generation with optional audio input and multi-shot narration support
  • Video Continuation: Extend existing clips seamlessly
  • Reference-to-Video: Up to five real-person reference inputs with vocal timbre matching
  • Video Editing: Text-prompted edits, reference image guidance, and style transfer

The update also introduces 3x3 grid-based image generation for batch visual output.

Why It Matters

ComfyUI has become the central hub for creative AI workflows, and adding Wan2.7 as a native node integration eliminates the friction of running separate tools. The reference-to-video capability with up to five person inputs and vocal timbre matching is particularly notable for creators producing character-driven content, as it allows consistent identity across generated clips without manual compositing.

The breadth of supported modes in a single integration is unusual. Most video generation tools specialize in one or two workflows. Having image-to-video, text-to-video, editing, continuation, and reference-based generation all accessible through ComfyUI's node graph means creators can chain these capabilities together in ways the standalone tools do not support.

Key Details

  • Minimum version: ComfyUI v0.18.5 is required for the Wan2.7 Partner Nodes
  • Audio integration: Both image-to-video and text-to-video modes support audio-driven generation, enabling lip-sync and sound-reactive animation
  • Multi-shot narration: Text-to-video supports scripted multi-shot sequences from a single prompt
  • Grid generation: 3x3 grid-based image output for rapid visual exploration
  • Open ecosystem: The Partner Nodes system allows third-party models like Wan2.7 to integrate without forking or modifying ComfyUI core

What to Do Next

Update to ComfyUI v0.18.5 through the standard update process, then install the Wan2.7 Partner Nodes from the node manager. Start with the image-to-video workflow using a single reference frame to test your hardware capabilities before moving to the more resource-intensive reference-to-video and multi-shot narration features.