Voice dictation and meeting transcription, powered by AI running entirely on your computer. Type with your voice in any app — emails, code editors, documents, Slack. Record calls with automatic speaker identification. No cloud. No accounts. No exceptions.
Download for MacClass action lawsuit for recording without consent. Lawyers using Otter risk waiving attorney-client privilege the moment a transcript hits their servers.
Sends transcripts to cloud AI and trains on your data. A therapist's session notes become training data. That is a HIPAA violation waiting to happen.
Consultants' strategy sessions, HR disciplinary meetings, board discussions — processed, stored, and monetised on servers you do not control.
Type with your voice in any app — emails, code editors, documents, Slack. Five formatting modes auto-switch based on context. Draft emails by speaking, vibe-code with voice, or provide context to AI tools — all hands-free.
Record calls with automatic speaker identification — who said what, with timestamps. Client calls, board meetings, depositions, therapy sessions — any Zoom, Teams, or Meet call. No bots join your call, no audio leaves your device.
Searchable transcripts with summaries and action items. Export to Markdown, PDF, or auto-save to your Obsidian vault. Meeting notes and client call transcripts, all searchable in one place.
Type with your voice in any app — emails, code editors, documents, Slack. Casual for Slack, professional for email, code for VS Code. Switches automatically based on context. Draft emails, vibe-code, provide context to AI tools.
Record calls with automatic speaker identification — who said what, with timestamps. Identifies up to 10 speakers. Cross-meeting voice recognition learns who's who over time. Meeting notes and client calls in one place.
Your conversations are processed by AI models running entirely on your device. On-device speech recognition. On-device speaker identification. On-device text formatting. Your voice recordings never leave your computer. Ever.
Your voice recordings never leave your computer. Ever.
We don't have servers that store your data. We haven't built the infrastructure to collect it. We couldn't access your transcriptions even if we wanted to.
Your conversations are processed by AI models running entirely on your device.
No accounts. No cloud. No uploads. No exceptions.
Attorney-client privilege, protected by design. Depositions, client calls, and case notes transcribed locally.
Session notes stay on your device. HIPAA-friendly by architecture — no cloud, no risk.
Client financial data processed locally. Audit meetings and advisory calls with complete privacy.
Strategy sessions with complete discretion. Client conversations never leave your computer.
Disciplinary hearings, interviews, and performance reviews kept on your machine.
Source protection for confidential interviews. Record and transcribe without third-party exposure.
M&A discussions, board meetings, and strategic planning with total confidentiality.
IRB-governed interviews and focus groups processed entirely on-device.
Classified briefings, operational debriefs, and intelligence analysis transcribed entirely on-device. No network required.
Or start with a 14-day trial with all features — no credit card needed.
No credit card required. All Pro features for 14 days. Reverts to free tier automatically.
Cancel anytime. App reverts to free — nothing breaks, nothing is lost.
Your subscription funds ongoing model improvements and macOS compatibility. Independent, privacy-first development.
Why license keys? We use license keys instead of accounts to protect your privacy. No email, no login, no tracking. Just a key that unlocks Pro features on your device.
I switched from Otter after the lawsuit. Remark does everything I need and nothing leaves my laptop. That matters when you're in healthcare.— Clinical researcher, NHS
The dictation modes are surprisingly good. It writes casual in Slack and formal in email without me changing anything. And it's completely local.— Product manager, London
We needed transcription that our compliance team would approve. Remark was the only tool where the answer to 'what data leaves the device?' was genuinely nothing.— Head of IT, financial services
Remark's speech recognition, speaker identification, and text formatting all run locally on Apple Silicon. Your voice recordings, transcriptions, and meeting data never leave your device. Remark downloads speech recognition models on first use and checks for app updates via Sparkle — but your content never touches a server. After initial setup, Remark works entirely offline.
No. After the initial download and model setup, Remark works entirely offline. You can dictate and transcribe meetings on a plane, in a secure facility, or anywhere without an internet connection.
Remark processes everything on your device and never transmits data. There is no server-side data handling to audit, no Business Associate Agreement needed, and no third-party data processor in the chain. Your IT and compliance teams can verify this independently using any network monitor. While we don't hold a HIPAA certification (that applies to covered entities and their associates), Remark's architecture eliminates the data transmission risks that HIPAA addresses. For attorneys, the same principle protects privilege — no transcript ever reaches a third-party server.
Recordings and transcripts are saved as files on your computer, in a folder you choose. Remark doesn't have a database or cloud sync. Your files are yours — you can move them, back them up, or delete them like any other file on your device.
Yes. Remark captures system audio, so it works with any meeting platform — Zoom, Microsoft Teams, Google Meet, or any other app that plays audio through your computer's speakers. No browser extension or meeting bot required.
Crash reports are completely optional and off by default. If you choose to enable them, they contain only the error type, stack trace, your macOS version, and your Remark version. No audio, no transcripts, no meeting content. You can also report issues via email, where you see exactly what's being sent before you hit Send.
It's on the roadmap. Remark currently relies on Apple Silicon for local processing, which makes macOS the natural starting point. A Windows version will follow once we've identified equivalent local inference options for that platform.