← Back to Blog

ComfyUI + DaVinci Resolve: The Complete Setup Guide

Everything you need to get ComfyUI running as a local AI backend for video work — models, extensions, VRAM management, and wiring it to your Resolve workflow.

What You Need Before Starting

This guide assumes you have DaVinci Resolve installed (Free or Studio). For the AI rendering side you will need:

  • An NVIDIA GPU with at least 8GB VRAM (10GB+ recommended for 512px renders)
  • Python 3.10 or 3.11
  • Git
  • ~20GB of disk space for models

AMD GPUs are possible but require the DirectML or ROCm builds of ComfyUI. This guide focuses on NVIDIA/CUDA.

Step 1: Install ComfyUI

Clone the ComfyUI repository and install dependencies:

git clone https://github.com/comfyanonymous/ComfyUI.git cd ComfyUI pip install -r requirements.txt

On Windows, use the ComfyUI portable package if you prefer to avoid Python environment management. On macOS, Apple Silicon users can use the MPS backend — set --use-mps when launching.

Step 2: Download the Required Models

ShotLock's default workflows require:

  • Base model: A Stable Diffusion 1.5 or SDXL checkpoint. For video work, Realistic Vision v6 or DreamShaper 8 work well for the cinematic presets.
  • IP-Adapter model: ip-adapter_sd15.safetensors (for SD1.5) or ip-adapter_xl.safetensors (for SDXL) — place in ComfyUI/models/ipadapter/
  • CLIP Vision model: clip_vision_g.safetensors — place in ComfyUI/models/clip_vision/
  • ControlNet model: control_v11p_sd15_canny.pth or the lineart variant — place in ComfyUI/models/controlnet/

Step 3: Install ComfyUI-Manager and Required Nodes

Navigate to ComfyUI/custom_nodes/ and clone ComfyUI-Manager:

git clone https://github.com/ltdrdata/ComfyUI-Manager.git

Restart ComfyUI, then install these node packs via the Manager UI:

  • ComfyUI IPAdapter Plus
  • ComfyUI ControlNet Auxiliary Preprocessors

Step 4: Launch ComfyUI

From the ComfyUI directory:

python main.py --listen 127.0.0.1 --port 8188

ComfyUI should now be accessible at http://127.0.0.1:8188. Leave this running in the background while you use ShotLock.

Step 5: Install ShotLock

Run the ShotLock installer and point it at your ComfyUI URL:

bash install.sh --license-key YOUR_KEY

The installer creates a backend launcher at ~/.shotlock/start-backend.sh and installs the Resolve panel script. Open DaVinci Resolve → Workspace → Scripts → ShotLock to launch the panel.

VRAM Management Tips

  • Use 512×512 frame size for most work — SDXL at 1024px requires significantly more VRAM
  • Set --lowvram when launching ComfyUI if you have less than 8GB
  • Close other GPU-heavy applications (games, other AI tools) during render jobs
  • The cleanup worker in ShotLock clears temp frames after each job to free disk space