Turn raw footage into a first edit from a description
Cardboard can assemble an editable cut from footage and a plain-language brief, then let you fine-tune the result in chat or on the timeline.
- What they did
- They uploaded footage, a written brief, or both, and Cardboard generated an editable first pass. After that, they tightened the cut by chatting with the agent or adjusting the timeline. It could shorten openings, remove pauses, filler words and repeated takes, add captions, change framing to 9:16, and pull in B-roll. In one documentary demo, it also transcribed a reference film and copied its editing rhythm using archival, stock, and generated shots.
- How it went
- The demo documentary used 12m 36s of agent time and was reduced to 65 seconds. It contained 73 shots, including 29 generated ones, plus 13 narrated lines. Each shot remained its own clip, so they could swap it out individually.
- Tips
- If you want the agent to mimic an existing piece, give it a reference and ask for a shot-by-shot match. Watch the plan carefully before approving anything involving paid features, and check whether the plan depends on Creator-only tools like voice cloning, web-sourced B-roll, or 4K export.
Try it yourself with Gemini
I have [raw footage / a folder / a Google Drive link] for a [video length]-minute piece about [topic] in a [tone] style. Build an editable rough cut with the best clips in order, remove dead space, filler speech, and repeated takes, and suggest captions and any needed B-roll. If I give you a reference video, match its pacing and shot structure as closely as possible. Show me a shot list with timestamps, and pause for my approval before you make any changes that would delete, send, submit, publish, or pay for anything. Leave my original files untouched.
Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.
Get 3 agent uses like this every morning
With the day's top AI news. Free, and one click to unsubscribe.
Related agent uses
Turn a PDF into an interactive course
An agent read textbooks and PDFs, built lesson sections, filled in interactive blocks, and sent the course to a second agent for review.
Turn last week’s GitHub activity into a blog post
An automated pipeline turns recent code activity into a first-person weekly draft and stages it for review in Notion and DEV.
Make a 3D Blender model from an image
A developer used a coding agent with a local Blender skill to turn a generated image into a Blender file.
Draft Figma layouts with a command-line agent
A custom agent used figma-use to generate Figma design elements and reusable UI pieces from a terminal.
Automate repeated site content updates
A developer used an agent to turn one podcast-page update into a repeatable workflow for their Nuxt site.
Let an agent package one idea for multiple channels
A creator delegated script writing, thread drafting, and post repurposing to a CreaoAI agent.