Le Scribe Audio

Le Scribe Audio is a professional 2-pass AI audio and video transcription web platform published in France by LE SCRIBE AUDIO (35650 Le Rheu, France). Built for academic researchers, CSE and CSSCT works council secretaries, municipal administrations, legal teams, consultants, and journalists, it converts multi-speaker recordings into structured, ready-to-edit documents.

Unlike single-pass speech-to-text tools, Le Scribe Audio runs a 2-pass AI pipeline that refines speaker turns, punctuation, paragraph structure, and terminology. It cuts transcription and review time in half (2 to 3 hours of full review and formatting per audio hour vs 4 to 7 hours of manual typing from scratch).

Key features include a synced interactive browser audio/video player, global renaming or coding of speaker diarization labels (such as M. Martin, Mme Dubois, or Enquêté P1) prior to export, an on-demand 1-click AI summary (executive summary followed by a chrono-thematic breakdown), and a standard 100% editable Word (.docx) export. Files are stored in France on OVHcloud (AES-256) with 1-click deletion, and AI inference runs on Google Cloud Vertex AI (Europe region, Zero Data Retention). Optional human proofreading by two successive transcribers is available at 60 EUR HT per audio hour.