macOS
Windows
Linux01 · Dictation
Speak into any app.
A global hotkey records anywhere. Clean, formatted text lands in whatever app you're in: email, editor, chat, terminal. About two seconds, whatever the length.
Speak.It appears.Done.
Mr.Speech is the connected voice workspace. Dictate into any app on your desktop, turn meetings into decisions and tasks, and keep an AI notebook that ChatGPT and Claude can read and write. With a fully private mode that never touches the internet.
~2 sfrom stop to finished text
3-4×faster than typing
~100dictation languages, auto-detected
Freeduring the open beta
Scroll. Your voice does the rest
Dictation
You talk like a human.It writes like a pro.
Keep scrolling to clean this up.
What you said
uh hi mister chen um thanks for your patience the updated contract is attached uh could you send it back signed by thursday um actually no by friday thanks
New message
To: Mr. Chen
Hi Mr. Chen,
Thanks for your patience. The updated contract is attached.
Could you send it back signed by Friday?
Thanks
Fillers gone. Punctuation placed. And because you're in an email window, it writes like an email. Mr.Speech formats for the app you're in, automatically.
Works in every app
One hotkey. Any window that accepts text.
Mr.Speech types directly into your focused window, including your inbox, code editor, terminal, or chat app. No window switching, no copy-paste, and your clipboard is untouched.
Gmail
Slack
Notion
Linear
Cursor
VS Code
Figma
Outlook
WhatsApp
DiscordGmailSlackNotionLinearCursorVS CodeFigmaOutlookWhatsAppDiscord
ChromeTerminal
Obsidian
Arc
GitHubOutlookSlackNotionCursorLinearChromeTerminalObsidianArcGitHubOutlookSlackNotionCursorLinear The connected workspace
Five tools, always apart.One voice connects them.
A sentence you speak in a meeting becomes a task with an owner and a due date. Same app, zero copy-paste.
02 · Meetings
Every meeting becomes decisions.
Both sides of the call, speakers labelled. One click for a summary, action items and a follow-up draft, pushed straight onto a board.
03 · Notes
A notebook that connects itself.
Wiki-links, backlinks and a living graph, with search by meaning that runs on your machine, not in someone's cloud.
04 · Plans
Talk becomes tasks.
Kanban boards with checklists, labels, due dates and reminders. A finished meeting's action items become a board automatically.
05 · AI agent
Your AI can act on all of it.
Hold a hotkey and ask your own agent to search, write and file notes. Or connect ChatGPT, Claude and Cursor: they can read and write your notes, tasks and boards, live today.
See it work
Ninety seconds is all you need.
A real Mr.Speech session showing the hotkey held, voice input, and the finished text dropped directly into the focused app. No edits, no tricks.
Privacy
What you saystays yours.
Mr.Speech has a fully private mode: dictation, notes, search and speaker recognition run entirely on your machine. No account, no internet, nothing sent anywhere. And when you choose the cloud for extra accuracy, it runs on European infrastructure.
Every note exports as plain Markdown. If you ever leave, you take everything with you.
Fully private mode
- Dictation on your machine
- Notes and search on your machine
- Speaker recognition on your machine
- Works with the internet unplugged
Cloud, when you want it
- Runs on European infrastructure
- No ads. No data resale.
Trust & compliance
Processed in the EU
Made in Germany
Processed in the EUFully private mode
GDPR compliant
Made in Germany The numbers
Measured, not promised.
5voice tools in one app: dictation, meetings, notes, plans, AI
1app doing the work of four
0copy-paste from meeting to task list
10 minof talking, finished about two seconds after you stop
9writing styles, from casual chat to formal email
3platforms: macOS, Windows, Linux
Enough watching.
Try it on your own machine. Free during the beta.
Download · v1.6.6
Get Mr.Speech for Linux.
Free beta · Account first · Native installer · Auto-updates enabled.
Download for LinuxLightOther platforms
Local-first by defaultAuto-updates enabledNative desktop installersSign in once in the app
Released 2026-06-05 · What's new in v1.6.6 →
PricingFree for local dictation.
Free for local dictation.
Free cloud beta active now.
The desktop app is free forever for local use. During our active open beta, every user gets a complimentary Pro account with 100 min/mo of cloud transcription, 180 min/mo of premium meetings, and notes sync, completely free.
Most popular
Free Beta
Unlocked automatically on signupFree during active open beta
Full cloud capabilities (100 min/mo cloud transcription, 180 min/mo premium meetings, 150 AI actions, notes sync).
Download beta now- Unlimited local dictation (offline capable)
- 100 min / mo cloud transcriptionmanaged high-accuracy cloud services
- 180 min / mo premium real-time meetings
- 150 AI actions / mo (Meeting AI / Notes AI)
- Cross-device notes sync
- Priority support during beta
Beta access is currently active. Final paid plan details will be shown before any purchase flow.
Product FAQ
Useful details before you install.
Still have a question? Email us , and a real person will read your message.
Yes. Local dictation, notes, and meeting recording can run on your computer. Cloud features are optional.
Yes, completely. Out of the box, every feature runs on your device with no internet required. To run the speech-to-text model locally on your computer, a graphics card (GPU) with at least 1 GB of VRAM is required (a dedicated GPU is highly recommended for smooth, real-time performance). If your system doesn't have a dedicated graphics card, you can easily use our managed cloud features instead. The notes workspace and meeting recording always work 100% offline.
