Le Scribe Audio is a French B2B SaaS platform available on Web, Android, and iOS built for academic researchers, works councils, public administrations, legal professionals, and professional transcriptionists. Powered by a two-pass multimodal AI pipeline on Google Cloud Vertex AI (Europe region, strict Zero Data Retention with no model training on user files), it transcribes pre-recorded audio and video files up to 10 hours and 10 GB—including professional .ds2 and .dss dictaphones, .mp3, .wav, and .mp4—without intrusive meeting bots joining calls. Users can switch between strict verbatim and cleaned full-text modes, globally rename speaker diarization labels inside a word-synced audio player, generate an executive summary followed by a chrono-thematic breakdown, and export a 100% editable Microsoft Word (.docx) document that cuts human editing and drafting time in half (2 to 3 hours of review per hour of audio instead of 4 to 7 hours of manual typing). All files are encrypted in France on OVHcloud (AES-256) with immediate one-click permanent deletion. Transparent pricing includes a 60-minute free trial without credit card, lifetime prepaid hour packs, and optional human proofreading at €60 excl. VAT/hour by two successive transcribers.