Video & audio tools

These nodes run on the server with ffmpeg. They do not call a generative model, but they still need the backend image to include ffmpeg and enough temporary disk for the input and output files.

Video tools

Extract Last Frame

Extracts the final decodable frame of a video and outputs a PNG image. Use it to continue a generated shot by connecting the frame to an image-to-video generator.

Thumbnail

Outputs a PNG frame. Set a non-negative timestamp for an exact position. Automatic mode first tries one second into the video to avoid a black intro, then falls back to the first frame for very short clips.

Merge Videos

Connect at least video1 and video2; video3 is optional. Clips are processed in port order and re-encoded to a common MP4 format before concatenation.

Video Trim

Keeps the interval from start to end in seconds. End must be greater than start.

Attach Audio

Combines one video and one audio stream. The video frames are preserved when possible and the resulting audio track is encoded for MP4 playback.

Speed Control

Changes video and audio speed together. Supported range is 0.25× to 4×; the server clamps values outside the range.

Caption Burn-in

Requires a video and either connected text or text entered in node settings. Configure font, size, weight, foreground colour, box or outline style, alignment, position and safe-area margins. Plain text wraps automatically and stays visible for the whole clip; SRT timing is preserved. Captions become part of the pixels and cannot be disabled in a player.

Audio tools

Audio Mix

Mixes one to three connected audio inputs. Empty ports are ignored. Connect separate voice, music or effects tracks when they should play together at full level.

Voice Ducking

Takes a voice track and a music track. Side-chain compression lowers the music while speech is present. Threshold controls when ducking starts; ratio controls how strongly music is reduced.

Audio Enhance

Applies ffmpeg noise reduction. NR controls reduction strength and NF is the noise floor in dB. Aggressive settings can introduce metallic artefacts.

Audio Trim

Keeps a start/end interval and can add fade-in and fade-out durations. Fade values are limited by the selected clip interval.

Batches

Video/audio tools fan out over iterator streams. Two multi-value inputs such as 3 videos and 2 audio tracks create a Cartesian batch, capped by the workflow's Max Combos setting.

Temporary files are deleted after each operation. If ffmpeg fails, the node reports error; if a required upstream stream is empty, it reports blocked.