LC AV Pipe Out π¬
Unpack the AV pipe into the individual wires your graph actually needs
- pipe
- av_pipe
- av_pipe
- Video
- Audio
- frame_rate
- duration
- sample_rate
- Width
- Height
- frame_count
- Positive prompt
- Negative prompt
- Seed
- total_steps
- sampler_name
- Models
The flip side of the pipe. Where LC AV Pipe (in/edit) bundles your whole video job into one wire, LC AV Pipe Out unbundles it at the other end: feed it an LC_AV_PIPE and it hands back the individual outputs - Video, Audio, frame rate, duration, resolution, frame count, and the generation metadata - so you can wire them to whatever needs a plain value or a plain IMAGE/AUDIO.
Why bother unpacking at all if the downstream Save Video nodes can read the pipe directly? Because not everything downstream speaks pipe. A preview node wants IMAGE, a text display wants a STRING, a resize wants numbers. The pipe is a great travel container and a poor interface for the last hop - this node is that last hop. It also passes the av_pipe through as its first output, so you can unpack and keep the bundle flowing in one node instead of choosing between the two.
How it works
This is the AV pipe's read end, and it has a second input that makes it more useful than a dumb unpacker: an optional pipe socket accepting the image pack's LC_PIPE. When both are wired, the image pipe's shared fields (prompts, seed, steps, size) are overlaid onto the AV pipe's values for the outputs - a handy way to pull generation metadata from your image-side graph into the AV side without a separate converter node. Note the source's framing: the LC_PIPE is an overlay, not an output - there's no LC_PIPE on the output side of this node, by design.
The outputs that matter
Ordered down the side, mirroring what was packed:
av_pipe- the bundle, passed through unchanged for further use.Video(IMAGE) andAudio(AUDIO) - the content, ready for preview, processing, or a Save Video's plain sockets.frame_rate,duration,sample_rate,Width,Height,frame_count- the numeric facts of the clip.Positive prompt,Negative prompt,Seed,total_steps,sampler_name,Models- metadata, for display or for feeding a save-metadata node.
How to install it
Part of lonecatone23/ComfyUI_LC_AV_nodes (LC Audio_Video nodes, MIT, by lonecatone23 - sibling to ComfyUI_LC123_nodes; install as its own repo):
- ComfyUI Manager: search "LC Audio_Video" or
ComfyUI_LC_AV_nodes. - Manual:
cd ComfyUI/custom_nodes && git clone https://github.com/lonecatone23/ComfyUI_LC_AV_nodes, restart ComfyUI ([LC AV] total 16 nodeson console).
No downloads or extra packages.
Where people trip
If an output looks empty, check where that slot was last written upstream, not this node - an unpacker can only surface what the pipe actually holds, and an empty slot means nobody packed it (or it was overwritten). And don't hunt for an LC_PIPE output here; that's the image pack's territory and this pack deliberately keeps the two types separate. If you genuinely need to go backward from AV pipe to image pipe, there's no dedicated node - convert the shared fields manually or design the graph to keep the LC_PIPE upstream.
Inputs (2)
| Name | Type | Default | Description |
|---|---|---|---|
| pipeopt | LC_PIPE | LC123 LC_PIPE. Overlays prompts/seed/steps/size. Not output. | |
| av_pipeopt | LC_AV_PIPE | LC_AV_PIPE to unpack. |
Outputs (15)
| Name | Type | Description |
|---|---|---|
| av_pipe | LC_AV_PIPE | β |
| Video | IMAGE | β |
| Audio | AUDIO | β |
| frame_rate | FLOAT | β |
| duration | FLOAT | β |
| sample_rate | INT | β |
| Width | INT | β |
| Height | INT | β |
| frame_count | INT | β |
| Positive prompt | STRING | β |
| Negative prompt | STRING | β |
| Seed | INT | β |
| total_steps | INT | β |
| sampler_name | STRING | β |
| Models | STRING | β |