Skip to content
WorkflowListedExternal
Open source

VACE First + Last Frame | Animate Keyframes with VACE! — Vace FL Native + Wrapper

VACE is incredibly flexible, this is yet another feature of it. You can even add control frames in as well (not shown in this workflow, but videos/workflows will be coming for that in the future). Check out the...

Sourcecivitai.com/models/1652752
TaskText to videoFamilyWanVersionEstimated VRAMNot yet characterizedBgrade

Runtime profile

Access
External workflow
Status
Listed
Version
VaceNativeFirstLast.json
Source
Civitai
Updated
Oct 18, 2026

Source description

VACE is incredibly flexible, this is yet another feature of it. You can even add control frames in as well (not shown in this workflow, but videos/workflows will be coming for that in the future). Check out the tutorial here if you need more guidance on how to use it!

What you need

Model files
0 GiB known subtotal

Remaining hardware or download-size estimates are not available yet.

Requirements

Models

4

  • native_umt5_xxl_fp8_e4m3fn_scaled.safetensors

    Possible

    Text encoder · Unknown

  • native_wan_2.1_vae.safetensors

    Possible

    VAE · Unknown

  • wan2.1_vace_1.3B_fp16.safetensors

    Possible

    Checkpoint · 135 B · Hugging Face · memescreamer/Wan_2.1_ComfyUI_repackaged

  • Wan21_CausVid_14B_T2V_lora_rank32.safetensors

    Possible

    LoRA · Unknown

Node packs

4

AI-generated commentary

AI-generated explanation based on source and configuration details. Suggestions are clearly labeled.

Purpose

  • This VACE workflow is for animating keyframes from first- and last-frame images plus text into a video.

Flow

  • Its graph includes VACE start-to-end-frame handling, sampling and decoding, image previews, and video combining.

Inputs and outputs

  • Image-loading and text-encoding nodes handle the inputs; image-preview and video-combine nodes handle the outputs.

Setup

  • The release contains two workflow files: VaceNativeFirstLast.json and VaceWrapperFirstLast.json.

  • The listed base model is Wan Video 14B t2v.

Dependencies

  • VaceNativeFirstLast.json lists four model files: native_umt5_xxl_fp8_e4m3fn_scaled.safetensors, native_wan_2.1_vae.safetensors, wan2.1_vace_1.3B_fp16.safetensors, and Wan21_CausVid_14B_T2V_lora_rank32.safetensors.

  • VaceWrapperFirstLast.json lists five model files: native_umt5_xxl_fp8_e4m3fn_scaled.safetensors, Wan2_1_VAE_fp32.safetensors, Wan2_1-VACE_module_14B_bf16.safetensors, wan2.1_t2v_14B_fp8_e4m3fn.safetensors, and Wan21_CausVid_14B_T2V_lora_rank32.safetensors.

  • The variants list these ComfyUI node packs: comfyui_essentials, comfyui-kjnodes, comfyui-videohelpersuite, ComfyUI-WanVideoWrapper, comfyui_layerstyle, and comfyui-easy-use.

Configuration

  • VACE can use control frames, but this workflow does not show how to set them up.

  • The description links a tutorial for more guidance.

Cautions

  • The source details list no model-use license.

Does this need to be edited?

Sign in to send an edit request.

Sources

1 source
CivitaiPrimaryVersion 1870737
civitai.com/models/1652752

Verified Oct 18

Source excerpts

3 excerpts
NoteOct 18civitai.com/models/1652752

Source context: 775 downloads · Type Workflows · Base model Wan Video 14B t2v

NoteAug 5civitai.com/models/1652752

Source context: 736 downloads · Type Workflows · Base model Wan Video 14B t2v

NoteJul 19civitai.com/models/1652752

Source context: 727 downloads · Type Workflows · Base model Wan Video 14B t2v