Speaking
120–166 WPM
Typical normal speaking rate ranges in speech-rate standards.
Native macOS menu bar app
Voice in. Text out. Kept local.
Susurr is private, system-wide dictation for macOS. Hold a shortcut, speak naturally, and your text lands in the field you are already using. No cloud upload. No subscription. No context switch.
Default shortcut
⌃
⌥
Space
Hold to dictate anywhere on your Mac.
WhisperKit on-device
macOS 14+ on Apple Silicon
Built for terminals, editors, and browser fields
Insert preview
Shortcut held
Terminal
Editor
Notes
active field launch-email.md
dictation inserted below
Draft the intro for the launch email and keep it short.
100% local transcription
Robust insertion fallback
Lightweight local history
One-time purchase
Built for real writing flow
Why voice feels faster
Your mouth moves quicker than your keyboard.
The research direction is consistent: people naturally speak much faster than they compose text on a keyboard. Susurr turns that speed advantage into usable text without pushing you through another app or dashboard.
Typing
19–33 WPM
Typical ranges for composing or transcribing text on a keyboard.
Direct text entry
~3x faster
Modern speech-entry studies show a meaningful speed advantage over typing.
Research note: ASA, Stanford HCI, Human Factors.
How it works
Three steps. Zero detours.
Susurr keeps the interaction small: trigger, speak, continue working. Everything else stays out of the way.
Press your shortcut
Use push-to-talk or toggle mode from the menu bar. It works system-wide, including in hard-to-reach apps.
Default
⌃
⌥
Space
Speak naturally
WhisperKit runs on-device, supports auto-detect and explicit language selection, and transcribes without sending audio anywhere.
Keep typing less
Susurr lightly cleans the text and inserts it back into the active field, with fallback paths for terminals, editors, and harder-to-reach app contexts.
Why people want it
Private dictation for actual work, not another AI everything suite.
Most speech tools fail on privacy, pricing, or insertion. Susurr focuses on one core promise: speak, and text appears exactly where you work.
100% local and private
No cloud upload, no subscription lock-in, and no awkward compromise for sensitive code, notes, or client content.
System-wide means everywhere
Use it in editors, email, browser fields, terminals, and IDEs. Susurr is built with explicit fallback paths for the places generic dictation usually breaks.
Fast enough to stay in flow
Hold the shortcut, speak, and keep moving. There is no context switching, copy-paste ritual, or extra dashboard to manage.
Mac-native by design
Menu bar first, lightweight, and built for Apple Silicon instead of shipping a heavy Electron shell.
One clear job
Susurr solves dictation and insertion directly. No meeting bot, no chatbot wrapper, no feature bloat.
Pay once
Buy it once, use it on three active Macs, and skip the monthly tool tax that piles up around writing work.
Privacy-first
Your words stay with you.
Susurr runs locally on Apple Silicon and uses a robust insertion strategy: accessibility when possible, clipboard restore when needed, and terminal-specific handling for the cases most apps ignore.
- No cloud dependency for transcription
- No recurring plan to keep core dictation unlocked
- Local history with retry and reinsert
- Custom cleanup rules and vocabulary
What changes with Susurr
Built for focused Mac users who care about privacy and reliable insertion.
Need
Typical tools
Susurr
Sensitive content
Often cloud-based
Local-only
Terminal insertion
Inconsistent
Explicit fallback strategy
Pricing model
Subscription-heavy
One-time purchase
Product scope
Feature sprawl
Single-purpose utility
Feature set
Small app, serious details.
The feature surface is intentionally compact, but the insertion logic and cleanup controls go deeper than the usual menu bar toy.
Native menu bar app
No Dock clutter. Susurr stays close, quiet, and ready.
Global shortcut
Hold or toggle push-to-talk across apps, terminals, and IDE internals.
WhisperKit transcription
Fast local speech-to-text with offline availability and strong accuracy.
Model and language choice
Select tiny, base, medium, or large models with multilingual auto-detect.
Text cleanup
Normalize whitespace, improve punctuation and capitalization, remove filler words, and apply custom vocabulary rules.
Insertion strategy
Accessibility write, clipboard restore, and iTerm-specific handling for better coverage everywhere.
Local history
Retry or reinsert recent transcripts without leaving the machine.
Simple onboarding
Guided setup for microphone access, accessibility permissions, and model download.
Model guide
Pick the model that fits your pace.
Susurr gives users a real speed-versus-quality choice. The practical default for most people is Whisper Small, with Tiny for minimal delay and Medium or Large V3 Turbo when accuracy matters most.
Best default
Recommended
Whisper Small
The best all-round choice for most users. Strong accuracy, solid speed, and one of the best current options for auto-detect.
Fast
Whisper Tiny
The lowest-latency option when fast response matters more than perfect transcription. Best for quick commands and minimal delay.
Best accuracy
Whisper Medium
A strong quality-focused choice when you want better transcription detail and can accept more compute cost than Small.
German-ready
Use explicit language
Whisper Large V3 Turbo
Excellent when quality is the priority, especially for German, but better used with the language set explicitly instead of relying on auto-detect.
English only
Base or Small (English)
A leaner option for users who only dictate in English and want a smaller footprint without exposing the multilingual models first.
Auto-detect notes
What not to use as the default
Do not make Base (English), Small (English), or Distil Large V3 the auto-detect default. Use Tiny and Large V3 Turbo with more caution when language detection matters.
Built for
People with high writing volume and low tolerance for tool noise.
Developers
Dictate commands, commit notes, and scratch ideas in terminals or IntelliJ without breaking your keyboard-first flow.
Founders and solo builders
Capture fast notes in tools like Slack, Notion, or mail while jumping between product, support, and execution.
Writers and marketers
Draft longer passages with less strain and fewer interruptions from mechanical typing.
Privacy-first power users
Keep private content local, avoid subscriptions, and use a focused utility instead of a cloud workflow dependency.
Pricing
Pay once. Use it forever.
€19.99 one-time for a lifetime license on 3 active Macs, plus applicable taxes at checkout.
Launch deal: the first 100 buyers save €4 with EARLYBIRD. That brings the discounted base price to €15.99 before any applicable taxes.
Lifetime license
3 active Macs
No subscription
Direct download
First 100 buyers save €4
€19.99 €15.99
Discounted base price before any applicable taxes. Then back to a €19.99 base price.
Get Susurr for MacSecure checkout via Polar. Taxes are calculated at checkout based on location. Lifetime license, 3 active Macs, no subscription.
FAQ
Questions that matter before you buy.
Susurr is built to work in terminals and IntelliJ using accessibility first, plus clipboard restore and terminal-specific fallbacks when needed. Exact behavior can still vary by app and focus context.
No. Susurr is designed as a local-first app. Transcription runs on-device with WhisperKit.
Susurr is built for Apple Silicon Macs running macOS 14 or newer.
No. The pricing model is intentionally simple: one purchase, lifetime use, three active Macs. Any applicable taxes are added in the Polar checkout.
A one-time Susurr license for 3 active Macs, delivered as a direct download with a Polar license key. Taxes, if applicable, are shown in checkout.