Video Transform
Deterministic, non-AI video operations exposed as one-call endpoints — resize, trim, merge, reverse, mirror, extract frames or audio, convert format, build contact sheets, convert to GIF. ffmpeg under the hood, REST on top. Predictable cost, identical output every time.
Pull the first and last frames out of a video as JPG stills — a fast way to thumbnail clips or seed img2img references.
Sample frames from a video at a configurable rate and return them in a ZIP. Standard prep step for frame-by-frame video AI workflows.
Convert a video clip to an animated GIF with a two-pass palette for clean colour. Configurable frame rate, width and duration.
Trim a video to a specific time range using stream copy — no re-encoding, no quality loss. Up to one-hour clips per call.
Reverse the playback direction of a video. Audio is reversed in step with the picture — useful for ping-pong loops and creative edits.
Concatenate two or more videos into one. Mixed resolutions, codecs and frame rates are normalised to the first clip's parameters.
Resize a video by width, height, or both. Aspect ratio is preserved when only one dimension is provided.
Flip a video horizontally or vertically. Useful for mirrored product showcases and symmetric kaleidoscope edits.
Convert a video between mp4, webm, avi, mov and mkv. Uses stream copy when the codec already fits the target container.
Apply a single visual filter to every frame of a video — mirror, grayscale, blur, sharpen, invert, or vertical flip. Lightweight post-processing for creative edits.
Generate a visual contact sheet from a video — evenly sampled frames arranged into a grid with optional timestamp or index labels.