Lectr

Lectr turns synced Apple Voice Memos into organized, searchable lecture transcripts without sending recordings to a server. A schedule-aware watcher routes each memo into its course folder, local mlx-whisper transcription produces part files, and Lectr combines them into one dated transcript per lecture.

I built Lectr as a single Go binary with an interactive Bubble Tea interface and one configuration file. It includes guided setup, dry runs, course/date/recording filters, shell completions, themed terminal output, and a launchd watcher with status and diagnostics. The local-first design keeps audio private while using Apple silicon acceleration to make the workflow practical day to day.