Video Audio Concat from URL

The name says "concat," but this isn't joining two videos end-to-end - it's muxing, replacing or attaching an audio track onto a video file. That distinction matters if you clicked in expecting something that stitches two clips together sequentially; for that, look at this pack's video_dir_combiner instead.

Why you'd reach for it

No video diffusion model produces audio - Wan, AnimateDiff, and everything else in that category outputs silent frames. If you've generated a clip in ComfyUI and separately produced a voiceover, a music bed, or a lip-sync track (through a TTS node, an audio pipeline, or a completely separate tool), this is the node that bolts the two together into a single playable file, without re-encoding the video itself.

The inputs and outputs that matter

• video_path - the source video.
• audio_source_path - the audio track to attach.
• output_dir - where the muxed result gets written.
• output_filename - default output.mp4.

One output: output_fullpath - the STRING path to the finished file.

The "from URL" in the display name suggests both path fields might accept remote http(s) URLs as well as local file paths - but the schema itself just types them as plain STRING with no indication of fetch behavior one way or the other, so I can't confirm that from the schema alone. If a remote URL is your actual use case, test it directly before relying on it; a local path is the safe, unambiguous option either way.

Installing it

ComfyUI Manager: search comfyui-huyl2-nodes, install, restart. By hand:

cd ComfyUI/custom_nodes
git clone https://github.com/huyl3-cpu/comfyui-sortlist

restart. Not documented in the pack's README. Muxing audio onto video without re-encoding is a textbook ffmpeg job (-c:v copy -map 0:v -map 1:a and similar), so expect a dependency on ffmpeg being available on PATH - if you're not running inside the author's preconfigured Colab environment (which typically ships ffmpeg preinstalled), you may need to install it yourself.

Common issues & troubleshooting

Mismatched durations are the most likely source of confusing output. Every audio/video muxing tool has to make a decision when the two inputs don't match in length - trim to the shorter, pad or loop to the longer, or just mux and let the player figure it out. The schema gives no indication which behavior this node uses, so check your first real output's duration against both source lengths before assuming either one drove the result.

If the node fails outright rather than producing a mismatched-but-playable file, an ffmpeg-not-found error (or an equivalent Python traceback about a missing binary) is the first thing to rule out, particularly outside a Colab environment where the author's own testing likely happened.