Hey y'all!
I'm very excited to announce the launch of Cutting Room AI, a natural language video editing copilot. Here's what it is...
Cutting Room AI is a standalone Windows desktop app that lets DaVinci Resolve Studio users control their timeline with plain English. Type what you want — "set opacity to 50% on all clips on track 2" or "add a red marker at the current timecode" — and the AI generates and executes the scripting API calls against your live Resolve session. No scripting knowledge required.
Key Features
Natural language commands executed against your live DaVinci Resolve timeline
Clip properties: opacity, zoom, pan, rotation, crop, composite mode, retime
Track operations: add, delete, enable/disable, lock, rename
Markers, clip colors, flags, media pool queries, and rendering control
Sandboxed script execution with AST-level validation and restricted builtins
Prompt library with pre-written commands to get started fast
Your feedback on the product will be greatly valued and appreciated!
Top comments (2)
The sandboxed execution with AST-level validation is the right call. Executing generated code against a live timeline without that guardrail is a recipe for destroyed projects.
One piece of feedback on the feature set. The clip properties you support (opacity, zoom, pan, etc.) cover most batch operations. But the gap I see is timeline navigation. Select all clips after the playhead or move this clip to track 3 would save more time than property adjustments for actual editing workflows.
Also curious about the prompt library. Are those community-contributed or curated by you? A shared library that updates based on usage patterns would compound value over time.
Thanks for engaging with my post! Yeah sandboxed execution was a must. Without it an 'AI deleted my project' situation was bound to arise, and there's only so much ctrl - z can fix.
Timeline navigation is a great call. You're right that "select all clips after playhead" or "move to track N" is where the real time savings live — property adjustments are useful but thats only a part of the equation. That's definitely on the roadmap now.
Right now the prompt library is curated by me, but a community library would be wicked. My experience with video editing is limited so evolving based on the language of professional video editors is exactly where I want to go.
If you're editing in Resolve and end up using Cry, I'd love to hear what workflows you throw at it first!