[Media] Building a video editing prototype in Rust using GPUI and wgpu

Hi, I've been experimenting with a video editing (NLE) prototype written in Rust.

The idea I'm exploring is prompt-based editing. Instead of manually scrubbing the timeline to find silence, I can type something like:

help me cut silence part

or

help me cut silence part -14db

and it analyzes the timeline and removes silent sections automatically.

I'm mostly editing interview-style and knowledge-based videos, so the goal is to see if this kind of workflow can speed up rough cuts in an NLE.

I'm also experimenting with things like:

cut similar subtitle (remove repeated subtitles)
cut subtitle space (remove gaps where nobody is speaking)

Another idea I'm testing is B-roll suggestions using an LLM.

The project is built with Rust using GPUI for the UI and wgpu for effect rendering, gstreamer and ffmpeg for preview and export. I'm still exploring the architecture and performance tradeoffs, especially around timeline processing and NLE-style editing operations.

It's still early and experimental, but I'm planning to open source it once the structure is a bit cleaner.

Curious if anyone here has worked on NLEs or media tools in Rust, or has thoughts about using Rust for this kind of workload.