Cutroom is a complete video editor with the modern AI tools built into the same timeline. Cut and arrange clips, caption from a local transcript, grade the colour, clean up audio, reframe for vertical, remove a background without a green screen, upscale, and export to H.264, H.265, or AV1. Do it all by hand, or ask your own AI agent to do it over MCP.
Reframe it for vertical, upscale it, and export AV1.
The same tools you click by hand. All on your machine.
Trim, split, speed ramps, transitions, transforms, masks, picture-in-picture, layers, and tracks. The editor you expect, all here.
Transcribe on your machine, then burn in captions with styles, animation, emphasis, and text placed behind the subject.
Grade the colour, level exposure, match shots, then denoise, normalise loudness, duck music under a voice, and remove silences.
Detect and track the subject, reframe for vertical, remove a background with no green screen, and upscale frame by frame.
Each of these is a tool in the editor, and each is also a tool the AI can call over MCP. So you can grade a shot by dragging sliders, or ask your agent to warm it up and match the next clip. The result is the same edit on the same timeline. There is no separate AI mode with a smaller set of powers.
A sample of the toolset
When you want new material, Cutroom can generate an image, a video, or a voice through the provider whose key you brought, on your own account. Record your screen or camera right into the timeline. Then export to H.264, H.265, or AV1, with hardware encoding when your machine has a supported GPU and a software fallback when it does not.
Generation goes to your provider
The models are yours; the editor is local.
Cutroom does not bundle a metered AI. Generation runs on the provider account you bring, and everything else, the cutting and the export, happens on your machine.
It is a full video editor: cut and arrange clips, speed ramps, transitions, layers, and picture-in-picture; transcribe and caption; grade colour and level exposure; denoise, normalise loudness, duck music, and remove silences; detect and track a subject, reframe for vertical, remove backgrounds, and upscale; generate images, video, and voice through your own provider; record screen and camera; and export to H.264, H.265, or AV1.
Yes. Every feature is a tool in the editor you can use directly. The AI is optional; it just calls the same tools when you would rather describe an edit than click it.
No. Transcription runs on your own machine, on the CPU, and captions are generated from that local transcript. Nothing is uploaded to caption a video.
Cutroom uses small on-device models to detect the subject and cut it out, so you can composite over another track without a green screen. It runs on your machine, on your frames.
H.264, H.265, and AV1. Cutroom uses your GPU's hardware encoder when you have a supported card (NVIDIA, AMD, or Intel) for speed, and falls back to a software encoder when you do not, so an export always finishes.
No. Generation runs through the provider whose key you brought, on your own account, at their rate. Cutroom adds no fee and does not resell tokens.
Detect and track the subject, reframe for vertical, remove a background with no green screen, and upscale frame by frame.