zambrini/slatecut

SlateCut is a desktop video editor built for people who film talking head or interview style videos, such as YouTubers or podcasters. Instead of editing by scrubbing through a timeline, it lets you edit based on the transcript of what was said: it can automatically cut out filler words and dead silence, use AI to pick the best take when you recorded a scene multiple times, and add zoom effects and captions. All video processing and rendering happens on your own computer using your hardware, and nothing you record is sent anywhere unless you turn on an optional cloud transcription or AI service yourself. The project is still in an early, actively changing stage, and the maintainer is upfront that it is rough around the edges, with breaking changes possible between updates. So far it has only actually been built and tested on Mac computers, both Apple Silicon and Intel chips. Support for Windows and Linux is planned in the design but has not yet been verified to actually work. There are no ready-made downloads yet, so using it means building it yourself from the source code. Under the hood, the app combines three pieces: a visual interface built with React, a core written in Rust that handles the actual video editing logic and talks to the interface, and a separate Python service running in the background that handles transcription and detecting things like silences, filler words, and good takes. These three pieces only run on your own machine and talk to each other locally. Getting it running requires installing several development tools, including Rust, Node.js, Python, and the video tool ffmpeg, then cloning the repository and installing dependencies for each of the three pieces before launching the app in developer mode. Optional cloud services for transcription or AI assistance can be added later by filling in an environment configuration file with the relevant service's access key, but none of that is required to use the core editing features locally.