Skip to content

Deploy WAN 2.x on RunPod

Use this page to choose a WAN video template, deploy it on a suitable NVIDIA GPU and open ComfyUI. The four templates cover text-to-video, long image-to-video, character animation and motion transfer.

Quick recommendation

  • Choose WAN 2.2 T2V to generate a video from a text prompt.
  • Choose WAN 2.2 I2V + I2LV SVI 2.0 PRO to animate a still image and extend it into a longer sequence.
  • Choose WAN 2.2 Animate for character animation or replacement driven by reference media.

  • Choose SCAIL-2 VI2V to transfer motion from a driving video to a reference character. Its FP8 profile is the documented 24 GB VRAM option.

  • Choose a template.

  • Check its hardware and storage requirements.
  • Open the RunPod deployment link and configure the pod.
  • Wait until the container log reports that the pod is ready.
  • Open Connect → HTTP Services → ComfyUI (port 8188).

Choose a template

Template Input Best suited to RunPod
WAN 2.2 T2V LightX2V Text prompt Generating a video directly from a description Deploy WAN T2V →
WAN 2.2 I2V + I2LV SVI 2.0 PRO LightX2V Image and prompt Image-to-video and longer videos assembled in chunks Deploy WAN I2V + SVI →
WAN 2.2 Animate LightX2V Character reference and driving media Character animation and character replacement Deploy WAN Animate →
WAN SCAIL-2 VI2V LightX2V Character image, driving video and optional prompt End-to-end motion transfer, including character replacement Deploy SCAIL-2 →

After deployment, open the ComfyUI workflow browser and select an example included with the chosen template.

Before you deploy

WAN workflows are sensitive to video length, resolution, chunk count and offloading. Choose the complete GPU and system-RAM profile rather than selecting a GPU based on VRAM alone.

Workload Tested GPU Minimum VRAM Minimum system RAM Notes
WAN 2.2 T2V or I2V, native ComfyUI L40S or RTX 6000 Ada 45 GB 90 GB Tested at 1024×768 and 122 frames
WAN 2.2 T2V or I2V, WanVideoWrapper L40S or RTX 6000 Ada 45 GB 50 GB Tested with offload 10
SVI 2.0 PRO long video L40S 45 GB 95 GB Tested at 1072×720 with 7 × 81-frame chunks
WAN 2.2 T2V or I2V with limitations A40 45 GB 45 GB WanVideoWrapper only; reduce resolution or duration
WAN 2.2 Animate L40S or RTX 6000 Ada 40 GB 105 GB Tested at 1024×768 with 77 chunks per sampler
SCAIL/SCAIL-2 FP16 L40S or RTX 6000 Ada 48 GB 96 GB Requirements vary with source length and size
SCAIL-2 FP8 RTX A5000 or RTX 4090 24 GB 60 GB Lower-VRAM SCAIL-2 profile

These values are tested minimums for the documented settings. Longer videos, additional chunks, higher resolutions and reduced offloading can require more VRAM and system RAM. See the full WAN hardware requirements.

Storage checklist

  • Allocate at least 90 GB of persistent volume storage mounted at /workspace.
  • Keep at least 15 GB of pod storage.
  • Allow additional workspace capacity for input videos, previews, chunks and generated output.
  • Do not reduce the template defaults unless you know the complete model footprint.

Optional credentials

  • Set PASSWORD if you want a fixed Code-Server password.
  • Add HF_TOKEN when downloading additional private or gated Hugging Face resources.
  • Add CIVITAI_TOKEN only when downloading resources from CivitAI.

See RunPod configuration for all supported environment variables and secret names.

Deploy and connect

  1. Open the deployment link for your chosen WAN workload.
  2. Choose an available GPU that matches a complete hardware profile above.
  3. Confirm the system RAM, pod storage and persistent /workspace volume.
  4. Add any required RunPod secrets or environment variables.
  5. Click Deploy Pod.
  6. Open the pod logs and wait for the final ready message. Initial startup can take several minutes while the container and models download.
  7. Expand the pod, select Connect, and open ComfyUI under HTTP Services.

For screenshots, normal startup checkpoints and troubleshooting, follow the full RunPod pod deployment guide.

Container date tags

RunPod templates use dated container tags. Keep the tag supplied by the template unless you are deliberately testing another build. Different WAN workloads can remain on different validated tags when a newer container has not yet been tested with every model and workflow combination.

Continue after deployment

Reference

Topic Guide
GPU, RAM and storage requirements WAN hardware requirements
Environment variables and network services RunPod configuration
Pod startup and troubleshooting RunPod pod deployment
Model documentation and upstream resources WAN resources