Cardboard

People who need to make product demos, launch videos, UGC ads, explainers, podcast clips and similar content often lack editing experience or time to manually cut dead air, trim retakes, source B-roll, write captions, or reframe footage for different formats. Today this work requires traditional timeline-based editors and manual searching for stock footage or archival clips, plus manual transcript cleanup to remove filler words and retakes. Teams also need to collect feedback and share drafts, which adds more steps. The page frames this as slow, skill-dependent work that keeps people without editing experience from producing polished video.

Cardboard lets users describe a video in a prompt, upload their own footage, or both, and generates an editable first cut on a timeline. It can generate video, images, music, sound effects and motion graphics from text, clone a user's voice for narration, source B-roll from the web or from linked videos, and match the pacing, color, or caption style of a reference clip. Users refine cuts through chat requests or direct timeline edits, including trimming dead air and filler words, pulling best moments from long recordings, adding captions, and reframing for formats like 9:16. Finished videos can be shared via link with comments or exported, including NLE exports, and the desktop app can connect to coding agents like Claude Code and Codex; rendering is described as about 12x faster than real time.