SeedVR2 AI Video Upscaling
ByteDance's SeedVR2 restores video resolution with strong temporal consistency using a diffusion system trained on video sequences (not frame-by-frame). Runs in ComfyUI on consumer GPUs with both GGUF-quantized and standard diffusion workflows.
✓ 100% free — download the workflow and run it locally.Preview
Requirements
- GPU: NVIDIA RTX 30XX/40XX/50XX (FP16)
- VRAM: 6 GB minimum (more recommended)
- OS: Windows
- Storage: 40 GB+ free
Model downloads
- Models auto-download via the SeedVR2 nodes (see custom nodes) https://github.com/numz/ComfyUI-SeedVR2_VideoUpscaler
Custom nodes
- ComfyUI Manager https://github.com/ltdrdata/ComfyUI-Manager
- ComfyUI-SeedVR2 VideoUpscaler (GGUF) https://github.com/lihaoyun6/ComfyUI-SeedVR2_VideoUpscaler
- ComfyUI-SeedVR2 VideoUpscaler (Standard) https://github.com/numz/ComfyUI-SeedVR2_VideoUpscaler
- rgthree-comfy https://github.com/rgthree/rgthree-comfy
- ComfyUI-VideoHelperSuite https://github.com/Kosinkadink/ComfyUI-VideoHelperSuite
Setup steps
- Install each custom node above and run its pip requirements.
- Reinstall PyTorch 2.6.0+cu124 if the nodes require it.
- Load a workflow below (Standard or GGUF) and feed it your video.
- Increase the blockswap value (max 36) if you run out of VRAM.
Notes & tips
- 480p upscale in 3–10 min on a 6GB RTX 4050; higher GPUs do 720p/24fps or 4K.
- Reduce batch size for memory issues (may add flicker).
Want the 1-click version?
Skip the manual model downloads and node installs — get a ready-to-run installer, or unlock all 75+ with Local Lab Pro.