Happy to answer questions about the architecture, the speaker ID pipeline, or anything else.

Known limitations right now: English only, macOS only (Apple Silicon), no team/sharing features.

I would love to hear what you'd use it for and what's missing.