Cycling through all six. Tap any point to stop.

Measured on six things

Two-pass AI audio & video transcription up to 10h into 100% editable Word (.docx) with OVHcloud France storage.

Growth

A real, specific wedge: GDPR-sensitive EU buyers (researchers, works councils, legal, medical) with dictaphone formats and France-hosted storage that Otter.ai and Rev don't lead with. Headroom is real but the segment is niche.

Screenshots

1 / 1

Le Scribe Audio

About this app

Le Scribe Audio (https://lescribeaudio.com/en) is a French B2B SaaS platform (Web, iOS & Android) built for researchers, works councils, public institutions, legal practitioners, and professional transcribers. Powered by a two-pass multimodal AI pipeline via Google Cloud Vertex AI (Europe region, Zero Data Retention), it transcribes single recordings up to 10 hours (10 GB)—including .ds2/.dss dictaphones, .mp3, .wav, and .mp4—without intrusive meeting bots. Users switch between verbatim and cleaned full-text modes, rename speaker labels globally in a synced audio player, generate an executive summary followed by a chrono-thematic breakdown, and export a 100% editable Microsoft Word (.docx) file that cuts human editing time in half (2–3h per audio hour). Stored in France on OVHcloud (AES-256) with 1-click deletion, 60-min free trial, lifetime hour packs, and optional certified human proofreading (€60 excl. VAT/h by two successive transcribers).

Key features

1Two-pass multimodal AI transcription via Google Cloud Vertex AI (EU region, Zero Data Retention), 100+ languages with bilingual code-switching.
2Single-file uploads up to 10 hours / 10 GB including .ds2/.dss dictaphone formats, MP3, WAV, MP4, MOV.
3Interactive synced player with 1-click global speaker renaming, plus Word (.docx), PDF, SRT and VTT exports.
4France-hosted storage on OVHcloud (AES-256), 1-click deletion, 60-minute free trial, lifetime hour packs from €2.50/h.
5Optional certified human proofreading at €60 excl. VAT/h by two successive transcribers.

Target customer profile

Drawn from the product itself, not from a survey.

Demographic

Researchers and academics

Pain points

Manual transcription is time-consuming and error-prone; strict data privacy requirements.

Primary needs

Accurate verbatim transcripts, GDPR-compliant storage, and export formats compatible with analysis software.

Demographic

Legal and medical professionals

Pain points

Need for precise terminology and confidentiality; compliance with legal and medical standards.

Primary needs

High accuracy, secure data handling, and support for dictaphone formats.

Demographic

Corporations and HR departments

Pain points

Efficient meeting minutes and interview transcripts; collaboration and quick turnaround.

Primary needs

Speaker identification, easy editing, and structured Word exports with optional summaries.

Scout Review

Written by AI from measured evidence, scored out of 100.

Le Scribe Audio is a French B2B transcription platform aimed squarely at buyers who can't send audio to a US cloud: researchers, works councils, lawyers, doctors. The pitch is specific and the site backs it up — OVHcloud France storage, Vertex AI EU with Zero Data Retention, .ds2/.dss dictaphone support, 10-hour single-file uploads, and a 100% editable Word export. The site is fast (Lighthouse 99) and the onboarding is genuinely low-friction (60 free minutes, no card). Where it's thin: documentation is a PDF, there's no SOC 2 or ISO badge, and the accessibility audit shows real defects. Against Otter.ai and Rev, the compliance story is the differentiator — everything else is table stakes.

Measure by measure

Usability

Clear 3-step flow, 60 free minutes with no card, and a live demo link lower friction than most peers. But the docs link resolves to a benchmark PDF, not a help center — support depth is unproven.

Design

Judging the marketing site: a dense but coherent modern SaaS layout — hero, live demo widget, 3-step flow, testimonials, pricing cards. Consistent card system and typography. App UI design could not be assessed (no store screenshots captured).

Speed

Lighthouse 99/100 on both mobile and desktop: LCP 1.2s, CLS 0.029, TBT 119ms. Genuinely fast site with no measured lag. Processing claims (~10 min per audio hour) are unverified but plausible.

Security

Strong header hygiene (HSTS, frame protection, Referrer-Policy, Permissions-Policy, SPF, DMARC; 10/12 checklist) plus France-hosted AES-256 and Vertex AI EU Zero Data Retention claims. No CSP, no SOC 2/ISO, no security page found.

Accessibility

Lighthouse a11y 92/100, but failing audits include buttons without accessible names and insufficient color contrast — real, fixable defects. 100+ language support is a plus; no WCAG statement or a11y page found.

Growth

Improvement areas

Documentation is a benchmark PDF, not a help center — no getting-started, FAQ, or API docs found.
No SOC 2, ISO 27001, or public security page despite targeting legal and medical buyers.
Three Lighthouse accessibility failures: unnamed buttons, low contrast, mismatched labels.
No CSP header; no public status page or incident history.

Design & Usability

The marketing site is a competent, modern SaaS build: hero with a live demo widget, a clear 3-step flow (upload → AI transcribe → listen/edit/export), testimonial cards from named researchers and lawyers, and pricing laid out plainly. Onboarding friction is low — 60 free minutes, no credit card, and a demo link right on the fold. The weak spot is depth: the docs URL resolves to a benchmark PDF rather than a help center, so anyone past the first hour is on their own. App UI design could not be assessed from the captured evidence.

Speed & Security

Speed is the standout: Lighthouse 99/100 on both mobile and desktop, LCP 1.2s, CLS 0.029, TBT 119ms — genuinely snappy for a page this dense. Security is solid-but-unproven: HTTPS enforced, HSTS, frame protection, Referrer-Policy, Permissions-Policy, SPF and DMARC all present (10/12 checklist), plus France-hosted AES-256 and Vertex AI EU Zero Data Retention claims. What's missing is the paperwork buyers in legal and medical will actually ask for: no CSP, no SOC 2, no ISO 27001, no security page.

Accessibility & Growth

Accessibility lands at 92/100 on Lighthouse but with real defects: buttons lacking accessible names, insufficient contrast, and visible labels that don't match accessible names — all fixable, none fixed. 100+ language support helps, but there's no WCAG statement. Growth is the more interesting story: the wedge is GDPR-sensitive EU buyers (researchers, works councils, legal, medical) plus dictaphone formats and 10-hour single-file uploads that Otter.ai and Rev don't lead with. That's a real differentiator against a real need — the limit is that the segment is niche and the incumbents are large.

Conclusion

If you're a French or EU institution that has been told 'no US cloud' by legal, this is a credible option with a clear pitch and a fair price (€1.50/h on the Pro plan, lifetime packs from €2.50/h). If you're an English-speaking team that just wants fast transcripts, Otter.ai is cheaper and more integrated. The gap between those two audiences is where this product lives — and it should lean harder into it with real certifications and real documentation.

Suggestions for improvement

Replace the docs PDF link with an actual help center: getting-started, export formats, dictaphone setup, billing FAQ.
Publish a security page with SOC 2 or ISO 27001 roadmap, subprocessor list, and a data-flow diagram — legal buyers will ask.
Fix the three Lighthouse a11y failures: accessible names on buttons, contrast ratios, and label/name matching.
Add a CSP header and a public status page; both are cheap and both are checked by procurement.

How we score

Alternatives comparison

Named competitors, point by point. Nobody paid to appear here or to be left out.

	Le Scribe Audio	Otter.ai	Rev
Primary positioning	EU/GDPR-first transcription for researchers, legal, medical, works councils	AI meeting notes with real-time collaboration and bot integration	AI + human transcription with fast turnaround, US-centric
Data residency	France (OVHcloud), Vertex AI EU region, Zero Data Retention	US-hosted cloud	US-hosted cloud
Dictaphone format support (.ds2/.dss)	Yes — explicitly supported	Not a headline feature	Accepted via upload, not a focus
Max single-file length	Up to 10 hours / 10 GB per file	Tier-limited; typically shorter per-file limits	Long files supported, priced per minute
Human proofreading option	Optional, €60 excl. VAT/h, two successive transcribers	Not offered	Core offering — human transcription is the product
Pricing model	€1.50/h Pro plan; lifetime packs €2.50–€3.75/h; 60 min free	Monthly subscription tiers with minute caps	Per-minute pricing, higher for human transcription
Meeting bot / live capture	No bots — upload-based workflow by design	Yes — OtterPilot joins Zoom/Meet/Teams	Limited; primarily upload-based
Published compliance certifications	GDPR claims, France hosting; no SOC 2/ISO found	SOC 2 and GDPR programs published	SOC 2 and HIPAA offerings published