Drop PAL into your ComfyUI graph. Spatial layout, camera, multipass render — beauty, depth, normal, alpha, ID matte — feeding straight into KSampler, ControlNet, or Video Combine. Renders locally on your GPU. Zero cloud tax.
Install the node, connect model_3d, render
a clean 512px beauty pass with full viewport + camera control —
no subscription required. Upgrade only when you need hi-res output, multipass, or Drive auto-save.
Five render passes out of one node. Drop depth straight into ControlNet, beauty into KSampler, alpha into your comp, ID matte for per-object masks.
Flip use_local_renderer=true and the node renders via wgpu (Metal / Vulkan / DX12) directly in Python. No iframe, no 15MB cap, nothing hitting our infra.
Spatial layout, camera rigs, lens presets, DOF, timeline — the whole PAL viewport embedded in your graph. Connect Load3D / Hunyuan3D / Tripo straight into the model_3d input.
Three.js loaders inside the viewport. Python-side, pygfx handles GLB/GLTF natively and assimp picks up FBX + everything else.
cd ~/ComfyUI/custom_nodes git clone https://github.com/Lenscowboy/comfyui-lenscowboy-pal.git pip install -r comfyui-lenscowboy-pal/requirements.txt # restart ComfyUI, search "PAL" in node picker
COMFY PAL is a node. If you need the full factory — script breakdown, multi-shot orchestration, continuity, client delivery — that's the LensCowboy SaaS pipeline.
See Enterprise →