Skip to main content

All Formats · MP3, MP4, WAV, M4A, MKV, OPUS & Dictaphones (.DS2)

Convert Audio & Video to Text: Up to 10 Continuous Hours Directly to Editable Word (.docx)

Whether you recorded an iPhone Voice Memo (M4A), an uncompressed field interview (WAV/FLAC), a full-day OBS or Teams video stream (MKV/MP4), or a legal/medical handheld dictaphone (.DS2/.DSS), convert any recording into a structured, speaker-labeled Microsoft Word document without hitting 200 MB upload walls.

Key Takeaways

  • Direct streaming upload of MP3, MP4, WAV, M4A, MOV, MKV, WEBM, FLAC, OPUS, and .DS2/.DSS up to 10 continuous hours and 10 GB.
  • 2-pass multimodal diarization across 100+ languages (including mid-sentence bilingual code-switching) plus 1-click global speaker-label coding.
  • Export to 100% editable Microsoft Word (.docx), .PDF, .SRT/.VTT subtitles, and CAQDAS / IRaMuTeQ text corpora.
  • OVHcloud France AES-256 storage + Google Cloud Vertex AI (Europe region, Zero Data Retention) + 60 free minutes without a credit card.

Video Overview (2 min 07): GDPR Audio Transcription in Europe Without Meeting Bots

See how European organizations, Works Councils (EWC), and research teams transcribe up to 10 continuous hours via our 2-pass architecture (AES-256 storage in France at OVHcloud, Google Cloud Vertex AI Europe with Zero Data Retention, and 100% editable Word export):

Interactive chapters (8) — click to jump
Test on your own recording — 60 free minutes, no credit card required

1. Supported Audio, Video & Dictaphone Formats at a Glance

Supported audio, video, and dictaphone input formats and output exports on Le Scribe Audio
Recording SourceAccepted ExtensionsMax Single UploadOutput Formats
Smartphones & Voice Notes.m4a, .mp3, .aac, .opus, .oggUp to 10 hours / 10 GB (4h on free trial)100% editable Word (.docx), PDF, IRaMuTeQ UTF-8
Field Recorders (Zoom, Tascam, Sony).wav (PCM 16–48 kHz), .flac, .aiffUp to 10 hours / 10 GB uninterruptedWord (.docx) with line numbers for NVivo/MAXQDA
Video Calls, Lectures & OBS Streams.mp4, .mov, .mkv, .webm, .aviAuto audio extraction up to 10 GBWord (.docx) + synced .SRT & .VTT subtitles
Legal & Medical Dictaphones.ds2, .dss (Olympus / OM System, Philips, Grundig)Direct upload without desktop codec toolsStructured Word (.docx) & PDF

2. Why Built-In Microsoft Word 365 & Single-Pass Converters Fall Short on Long Files

Microsoft Word for the Web caps uploaded audio at 200 MB per file and 300 minutes per month, rejecting 2-hour .mp4 recordings, uncompressed .wav sessions, and .ds2 dictations outright. Meanwhile, single-pass Whisper converters suffer from hallucination loops during room silence and scramble speaker identities whenever you manually chop a 3-hour conference into smaller pieces.

Le Scribe Audio's 2-pass multimodal engine reasons directly over the acoustic signal across the full recording, ranking #1 in our reproducible 10-software benchmark published on CNRS HAL (hal-05865820) and Zenodo (DOI 10.5281/zenodo.24215193) with 2.5x fewer Word Track Changes edits than the market average.

3. Interactive Verification Studio, Speaker Coding & Realistic Time Savings

  • Strict Verbatim or Full Clean Read: Choose strict word-for-word capture for qualitative research and hearings, or full Clean Read (which removes vocal hesitations while preserving 100% of spoken content without summarizing).
  • On-demand 1-click AI synthesis: Generate an Executive Summary followed by a timestamped Chrono-Thematic Breakdown ([hh:mm:ss - hh:mm:ss]) directly inside the interactive player — never speculative action-item lists.
  • 1-click speaker-label coding: Rename Speaker 1 and Speaker 2 globally to Mr. Bernard, Ms. Dubois, or Participant P1 across every turn (updating diarization speaker labels only; spoken proper nouns inside sentences remain under your editorial review).
  • Cut human review time in half (÷ 2): Instead of spending 4 to 7 hours typing 1 hour of audio manually from scratch (4–6h pro, 5–7h non-pro), full ear-checked proofreading and formatting in our synchronized player takes 2 to 3 hours per hour of audio before exporting your 100% editable Microsoft Word (.docx) file.

4. Pay-As-You-Go Lifetime Packs, Pro Team Rollover & GDPR Security

Start immediately with 60 free minutes (no credit card required). Scale with non-expiring lifetime packs (2h = €7.50, 5h = €15.00, 10h = €27.50, 25h / 3 seats = €65.00, 50h / 5 seats = €125.00, 240h Annual Pass / 5 seats = €360.00 excl. VAT), our Pro & Team subscription at €30.00 excl. VAT/month for 20h (€1.50/h, 5 seats, automatic rollover), or €60.00 excl. VAT/h dual-transcriber human proofreading for French recordings. Review our GDPR & OVHcloud France Security Architecture and Qualitative Research & IRaMuTeQ / MAXQDA Guide.

Frequently Asked Questions

Which audio, video, and dictaphone formats can I convert directly to text?

Le Scribe Audio natively accepts MP3, M4A (Apple Voice Memos), WAV, AIFF, AAC, FLAC, OGG, OPUS (WhatsApp/Signal voice notes), MP4, MOV, MKV (OBS Studio), AVI, WEBM, and encrypted professional dictaphone files (.DS2 and .DSS from Olympus / OM System, Philips, and Grundig) without prior file conversion.

Do I need to split multi-hour recordings or large video files before uploading?

No. You can upload continuous files up to 10 hours long and 10 GB in size (up to 4 hours during your 60-minute free trial) in a single block, with automatic chunked upload recovery on unstable Wi-Fi connections. Our 2-pass pipeline processes internal 20-minute windows with 5-second acoustic overlaps so zero words are lost at segment boundaries.

What export formats are included once my recording is transcribed?

Every transcription can be exported immediately as a 100% editable Microsoft Word (.docx) document (with speaker headings, optional timestamps, optional continuous margin line numbers, and a transparent AI attribution footer notice recommending human verification), a clean PDF, synchronized video subtitles (.SRT and .VTT), or UTF-8 text compatible with IRaMuTeQ, NVivo, MAXQDA, and ATLAS.ti.

How long does AI conversion take and how fast is full human proofreading?

Automated 2-pass AI processing takes roughly 10 minutes per hour of audio. If you perform a thorough start-to-finish human re-listening and formatting pass, expect 2 to 3 hours of work per hour of audio in our synchronized click-to-listen player — cutting your drafting time in half (-50%) compared to 4 to 7 hours of manual typing from scratch.

Ready to cut your transcription and drafting time in half?

Start immediately with 60 free transcription minutes — no credit card required, no mandatory subscription, and 100% editable Microsoft Word (.docx) export.