Pepys – Transcribe Anything

Pepys gives Claude transcription that goes beyond speech-to-text.

Transcribe audio, video, supported public media links or entire podcast feeds into accurate, speaker-labeled, timestamped transcripts in 99+ languages.

Claude can start a transcription, poll it to completion, retrieve structured transcript segments, generate summaries, export correctly timed SRT, VTT, TXT, Markdown or JSON, and search inside hours-long recordings without loading the full transcript into context.

Pepys also supports citation-backed transcript Q&A, so Claude can answer questions using verified claims tied back to the source transcript instead of relying on free-form summarization.

With the Pepys connector, Claude can:

• transcribe files and supported public links
• batch-transcribe RSS or Apple Podcasts feeds
• retrieve transcript segments, summaries and speaker labels
• search long transcripts for exact phrases, topics or moments
• ask questions and return citation-backed answers
• export SRT, VTT, TXT, Markdown and JSON
• check available transcription credits before starting large jobs

Authentication uses OAuth with the user’s own Pepys account, so the hosted connector requires no API key. Usage draws from that account’s Pepys credits. New accounts receive 60 free minutes, then transcription is pay-as-you-go rather than subscription-based.

Privacy: Pepys does not train on your audio or transcript content.

Built for long-context workflows: search hours of transcript, retrieve only the relevant sections, and keep Claude’s context focused on what actually matters.