Deep Transcribe

Example Use Cases

• 1

  Transcribe video or podcast audio into clean, formatted Markdown/HTML with speaker diarization and timestamps.

• 2

  An MCP-accessible transcription service (transcribe_annotated, transcribe_formatted, transcribe_basic, transcribe_deep) to integrate with desktop assistants or other automation.

• 3

  Enriched transcripts with section headings, summaries, frame captures, and optional research annotations for deeper content understanding and citation.

Description

Deep Transcribe downloads audio or video (e.g., from YouTube) and produces rich transcripts with speaker identification, timestamps, section headings, summaries, frame captures, and optional research annotations. It uses Deepgram for transcription/diarization and Claude or OpenAI for analysis and summarization, and writes outputs as Markdown and HTML with cached media and intermediates. The project can run as an MCP server exposing transcription actions (annotated, formatted, basic, deep) for integration with agents and desktop tools like Claude Desktop.