AD_video_merge

The author's own description is the best pitch this node needs: "merge videos with audio (like Jianying)." Jianying is ByteDance's wildly popular Chinese video-editing app (CapCut's sibling), and that's exactly the job here - take a bunch of separately generated clips and glue them into one video with synced audio, without leaving ComfyUI to do it in a real editor.

This matters more than it sounds like once you're actually doing video work. Generate a scene with Wan or LTX in 5-second chunks (which is how most local video models still work best), and you end up with a folder of disconnected clips instead of one video. AD_video_merge is the node that turns that folder back into a single deliverable, right at the end of your graph.

It takes up to 10 optional VIDEO inputs (video1 through video10 - plug in however many you actually have, leave the rest empty) and combines them according to merge_mode, which defaults to sequential - one after another, in input order. Three more required settings control the technical side of the join: target_fps (1–120, default 24 - the framerate every clip gets normalized to, since your source clips may not all match), audio_sample_rate (16000–48000 Hz, default 44100, the standard CD-quality rate), and force_audio_merge (boolean, default on) which tells the node to combine the audio tracks even if some clips are silent or mismatched rather than dropping audio on a mismatch. The single output is VIDEO - one merged file, ready to save or push further down your pipeline.

There isn't much to configure wrong here, which is the point. The three required fields are really just "what format should the output be" - pick an fps and sample rate that match your source material (or your target platform) and leave force_audio_merge on unless you specifically want silent clips to stay silent instead of getting a merged track.

Installing it: via ComfyUI Manager, search "ComfyUI-Apt_Preset" and install, or clone manually - cd ComfyUI/custom_nodes && git clone https://github.com/cardenluo/ComfyUI-Apt_Preset - then run install.bat to grab dependencies and restart ComfyUI. Video merge with audio muxing typically leans on ffmpeg under the hood for packs like this one; if you're on Linux/macOS and don't already have ffmpeg on your PATH, install it through your package manager before troubleshooting anything else.

Troubleshooting: if the merge fails or the output has no sound, the first thing to check is whether ffmpeg is actually available to ComfyUI's Python environment - that's the most common silent-failure cause for any video-plus-audio node, not just this one. If clips merge but look stuttery or the audio drifts out of sync, mismatched source framerates are the usual culprit - target_fps normalizes the output, but wildly different input framerates (say, mixing 16fps and 30fps clips) can still produce visible judder at the seams even after normalization, so it helps to generate your source clips at a consistent fps to begin with rather than relying on this node to smooth it all out.