I've been exploring video analysis through Qwen2.5-VL — not for editing but for understanding content. The model can describe what happens in videos frame by frame when you pass URLs or extracted frames. Not production-ready but the open-source vision models are getting surprisingly good at this.
What made you choose Rust over something like Tauri?