Transcribe the whole interview backlog on your Mac

Stringout transcribes interview footage on the Apple Silicon Mac holding your connected drives. On-device transcription is free and unlimited, so a growing backlog does not require purchased transcription hours. The recordings are not uploaded for that processing.

Keep the interviews in the same indexed library you will search and review later. In this example, field interviews sit beside expedition footage, Arctic wildlife and night skies, so the whole search for the northern lights stays in one place. Pro team sharing is a separate cloud service.

Search the interview transcripts for the phrase or topic, then open the matching recording at that passage. In the northern-lights example, a search for the first sighting brings you back to the speaker’s description of the sky. Play the surrounding answer to check the wording, delivery and context before treating it as a select.

Add the recording to a Stringout project and ask Auto-assemble to cut around that answer. Check the placed passage against its recording. It stays linked to the whole recording, so you can extend it in your NLE or come back for a different answer later.

We’ve adapted NVIDIA’s Parakeet speech recognition for Apple Core ML to run locally on Apple Silicon. Efficient transcription, word-level timestamps and native punctuation give interview editors readable dialogue tied to the source recording.

Your Mac does the processing, so your recordings stay on your drives and a growing backlog doesn’t become a bill for transcription hours.

NVIDIA Parakeet model collection