Skip to content

bridges.md

Latest commit

 

History

History
75 lines (56 loc) · 3.24 KB

bridges.md

File metadata and controls

75 lines (56 loc) · 3.24 KB
Raw

English | 简体中文 images/bridges.png

Bridge nodes — connecting other ComfyUI plugins

Any ComfyUI plugin that outputs IMAGE / VIDEO / AUDIO (mesh2motion, IPAdapter, ControlNet preprocessors, 3D nodes, any future plugin) can plug into a ComfyTV pipeline through the bridge nodes.

[any plugin]   IMAGE          [ComfyTV → Image]   COMFYTV_IMAGE        [Image Picker / Upscale / …]
  output    ──────────────→     bridge stage    ──────────────────→     ComfyTV stages
                              (Run + snapshot)

The bridge is itself a ComfyTV stage with its own Run button. Click Run, the upstream runs once, the URL is stored as the snapshot; downstream ComfyTV stages then use that snapshot on subsequent Runs instead of re-running.

Into-bridges

ComfyTV/Bridge contains 5 into-bridge nodes that take a native or third-party plugin's output and pull it into ComfyTV:

Node Input Output
→ ComfyTV Text STRING COMFYTV_TEXT
→ ComfyTV Image IMAGE COMFYTV_IMAGE
→ ComfyTV Images IMAGE batch COMFYTV_IMAGES
→ ComfyTV Video VIDEO COMFYTV_VIDEO
→ ComfyTV Audio AUDIO COMFYTV_AUDIO

Into-bridges have a Run; their output persists with the project.

Out-bridges

The reverse direction also exists: 5 out-bridge nodes turn a ComfyTV snapshot back into native ComfyUI types, so any plugin can consume what a ComfyTV pipeline produced:

Node Input Output
← ComfyTV Text COMFYTV_TEXT STRING
← ComfyTV Image COMFYTV_IMAGE IMAGE
← ComfyTV Mask COMFYTV_IMAGE MASK
← ComfyTV Video COMFYTV_VIDEO VIDEO
← ComfyTV Audio COMFYTV_AUDIO AUDIO

← ComfyTV Mask reads the image's alpha channel as a mask — handy for feeding a Cutout or painter result into native inpainting nodes.

Common patterns

Plugin output → ComfyTV downstream

[mesh2motion] ─VIDEO─→ [→ ComfyTV Video] ─COMFYTV_VIDEO─→ [Video Color / Stabilize / …] → …

Round trip: ComfyTV result → native plugin → back

[Image Stage] → [← ComfyTV Image] ─IMAGE─→ [any native plugin] ─IMAGE─→ [→ ComfyTV Image] → …

Plugin output is a multi-frame IMAGE batch

For plugins that output an IMAGE batch but no VIDEO object, chain ComfyUI's Create Video first:

[plugin] ─IMAGE batch─→ [Create Video (fps)] ─VIDEO─→ [→ ComfyTV Video] → …

Third-party LLM / prompt enhancer → ComfyTV prompt

Any ComfyUI node that outputs STRING (a prompt enhancer, a captioner, a custom LLM) can feed text into a ComfyTV pipeline:

[Prompt Enhance (Comfy-Org)] ─STRING─→ [→ ComfyTV Text] ─COMFYTV_TEXT─→ [Image / Video Stage] → …

File locations

Into-bridges write to ComfyUI's output directory under output/ComfyTV/bridge/…:

Bridge Format Subfolder
Image / Images PNG (one per frame) output/ComfyTV/bridge/
Video MP4 (auto codec) output/ComfyTV/bridge/
Audio WAV (universal, no codec deps) output/ComfyTV/bridge/