motionknit Back to home

Use case

Script to video, with real footage

Most script-to-video tools make a slideshow of stock clips that roughly match a keyword. MotionKnit reads the script the way an editor does: line by line, deciding what each line needs to show and holding each picture exactly as long as the voice is on it.

The script decides the cut

The script is divided into beats of a few words each, and every beat gets its own shot. Cuts fall on the words of your voiceover, measured from a word-level transcription — not on a fixed rhythm.

Every shot is judged against its line

For each beat the engine writes a search, pulls candidates from real video libraries and scores them frame by frame against what the line says, rejecting diagrams, talking heads and watermarked uploads. When no clip fits it uses a photograph.

You can change any of it

The output is an editable timeline, not a flat file. Swap a shot, retime a beat, rewrite a card or change the grade in the browser editor, then export. The project file is yours too.

Make a video with MotionKnit A script and a voiceover in, an editable video out.