listenin records microphone and system audio on macOS, then turns the recording into a high-quality, speaker-labeled transcript after the meeting. Audio and voiceprints stay on your machine, and retained recordings can be transcribed again as models and settings improve.
Features
- Local capture — records mic and system audio through macOS ScreenCaptureKit
- Batch transcription — uses long-form decoding and full-recording context instead of trading accuracy for live output
- Speaker diarization — clusters speakers across the whole recording and supports reusable, explicitly enrolled voice profiles
- Append-only passes — new transcriptions never overwrite old ones, making models and settings easy to compare
- Canonical YAML — stores structured speakers, timestamps, and segments, with Markdown as a regenerable view
- Glossary support — primes recognition with names, product terms, and company-specific language
- Optional refinement — cleans transcripts and produces summaries through your own OpenAI-compatible endpoint
- On-demand peek — transcribes the meeting so far when a quick mid-call check is useful