Built for the agents you already use
Edit video like a document
Delete a sentence from the transcript and the cut happens. No scrubbing, no hunting for the edit point.
Point at your words, get footage
Tag a phrase in your own transcript and the generated clip lands on the timeline already locked to it.
Ten times fewer clicks
Cutting filler words by hand means finding every one, selecting it, deleting it, then closing the gap. Saying it once does all of them.
Locate each filler, select the span, ripple delete, verify the join. Repeated 34 times.
“Cut every filler word, then tighten the pauses.” The agent resolves the spans and does the rest.
Review that never points at nothing
Comments anchor to a commit, a timestamp and a clip — so feedback survives the next edit instead of quietly rotting.
FAQ
No. It's a full manual NLE — timeline, trim modes, color wheels, keyframes. The agent is an operator you can hand the controls to, not a requirement.
Every edit is versioned automatically as it happens, so you can step back to any earlier state. Nothing the agent does is one-way.
Claude, ChatGPT and Grok are supported today over MCP. You add Arsaze as a custom connector, pick a project, and its tool calls play back live in your editor.
Not without your consent. Your uploads, generated video and voice models stay tied to your account — see the privacy policy for the full detail.
No. Rendering runs on our servers, so the editor is just a browser tab and your laptop never does the encoding.
Yes — export a finished MP4 whenever you want. You're never stuck inside the editor to get your video.