Voice-to-text for macOS

Voice in.
Text out.
Instant.

Press a hotkey, speak naturally, and watch your words insert instantly in any app. Local AI. No cloud required.

See Features

One-time purchase — use on 3 Macs, lifetime updates

Because paying a subscription for a Whisper wrapper is crazy.

Cursor
AI

How can I help you with your code today?

Me

Type a message...
Mail
To: Product Team
Subject: Weekly Sync Update

Hi everyone,


Notes
Q4 Roadmap Ideas
Edited just now
Vibe code with superpowers ⚡️
Global Hotkey

Push-to-Talk Dictation.

It's like a walkie-talkie for your Mac. Hold the shortcut, speak your mind, release to insert. Or tap once to toggle. Works in Xcode, Slack, Notes, browser — everywhere.

Hold to Speak

Sotto listens while you hold and transcribes when you release.

Instant Insert

No copying needed. Text appears directly in your active app.

Shortcuts
Cmd
+
Shift
+
Space
Space
Start / Stop Recording
CancelEsc
Typing
40 WPM
Speaking
150+ WPM
3.7x Faster
2.5hrs
Saved per week
100%
Flow State
Velocity

Stop typing at the
speed of slow.

The average person types at 40 words per minute but speaks at 150. Your brain works faster than your fingers. Sotto closes that gap.

Don't let friction kill your flow state. Capture complex thoughts, detailed feedback, and long-form content in seconds, not minutes.

Vibe Coding

Speed up your
vibe coding.

Let's be honest: when you're typing, you're lazy. You type "fix this" and hope the AI reads your mind. And then you get garbage code back.

With Sotto, you can effortlessly ramble about the edge cases, the architecture, and the specific constraints. Better context in = better code out.

  • Describe complex bugs in seconds
  • Explain architectural decisions clearly
  • Give detailed code review feedback
  • Stop wrestling with context windows
Cursor
Lazy Typing
User: fix the auth bug
AI: I'm not sure which bug you mean. Could you provide...
Sotto Voice
Transcribed instantly:"The auth middleware keeps 401ing because the refresh token logic in the interceptor isn't waiting for the new token to return before retrying the failed requests. We need to implement a queue there."
AI: Understood. I'll update the Axios interceptor to queue requests...
Works everywhere

Your voice, any app.

Sotto works as a system-wide layer. Talk anywhere, type everywhere.

Communication

Reply in seconds, not minutes.

Slack
🎨
Emma2:34 PM
Hey, can you review the new landing page design? Need your feedback before EOD
Tasks

File bugs while they're fresh.

Linear
SOTTONew Issue
Title
Description
Writing

Capture ideas at the speed of thought.

Notes
Dec 11, 2024 • Brain Dump

Figure out Q1 priorities

✓ Finalize pricing model

Works with Slack, Discord, Linear, Notion, email, and literally any app with a text field.

Voice-first. Privacy-first.

Built for professionals who value speed and privacy.

Push-to-Talk or Toggle

Hold to speak, or press once to start/stop. Fully customizable hotkeys that work system-wide.

Auto-Paste Anywhere

Transcripts paste directly into any app. No copy-paste needed. Just speak and it appears.

100% Local & Private

Default models run entirely on your Mac. Your audio never leaves your device.

Cloud Models Available

Connect OpenAI or Groq for maximum accuracy. Great for accents and technical terms.

Native Swift App

Built with Swift and SwiftUI for a true native macOS experience. Lightweight and fast.

Recording History

Every recording saved with timestamps. Search, replay, and re-transcribe anytime.

Import Audio

Drop audio files here

or click to browse

.mp3.m4a.wav.webm
Recent Imports
team-standup-dec4.m4a2:34
voice-memo-ideas.mp30:45
Import Audio

Transcribe any audio file.

Got voice memos, meeting recordings, or podcast episodes? Just drag and drop them into Sotto. We'll transcribe them using whichever model you choose.

Voice Memos
Meeting Recordings
Podcast Episodes
Any Audio File
Or use Cmd+Shift+I to import
Custom Vocabulary

Teach Sotto your words.

Add specialized terms, names, and jargon to your dictionary. Sotto uses these as hints for better accuracy on your unique vocabulary.

Technical Terms

Acronyms, frameworks, and industry jargon.

Names & Brands

People, companies, and products you mention often.

Foreign Words

Loanwords and phrases from other languages.

Dictionary

Your Vocabulary

12 words
SwiftUI
Whisper.cpp
OpenAI
macOS Sonoma
Raycast
TypeScript
Next.js
Vercel
Never Lose a Recording

Your audio is always saved.

Every recording is saved locally on your Mac. Not happy with the transcription? Re-transcribe with a different model anytime. Switch from Tiny to OpenAI with one click.

Re-transcribe Anytime

Try a different model if the first result wasn't perfect.

Play Back Audio

Listen to your original recording whenever you need.

Search Everything

Full-text search across all your transcripts.

Pro tip: Start with Tiny for speed, then re-transcribe important recordings with OpenAI for maximum accuracy.

Recording History
Today, 2:34 PM0:42
Tiny

Remember to update the API docs...

Today, 11:20 AM1:15
OpenAI

Meeting notes: discussed Q1 roadmap...

Yesterday0:23
Small

Quick reminder about the deploy...

Transcription Models

Choose your engine.

Run locally for privacy, or connect to the cloud for maximum accuracy. All models use the Neural Engine for fast inference.

WhisperKit Models

Private & Offline — Powered by Apple's Neural Engine

Tiny

Size~66 MB
SpeedFastest
AccuracyGood

Quick notes, drafts

Base (English)

Size~105 MB
SpeedFast
AccuracyBetter

Everyday use

Small (English)

Size~330 MB
SpeedFast
AccuracyGreat

Important documents

Recommended

Large V3 Turbo

Size~954 MB
SpeedFast
AccuracyExcellent

Best overall quality

Distil Large V3

Size~800 MB
SpeedFast
AccuracyExcellent

Great speed/accuracy balance

NVIDIA Parakeet Models

State-of-the-art accuracy — Powered by Apple's Neural Engine

Parakeet v2 (English)

NVIDIA Parakeet
Size2.6 GB
SpeedVery Fast
AccuracyBest

Highest recall, English only

Parakeet v3 (Multilingual)

NVIDIA Parakeet
Size2.7 GB
SpeedVery Fast
AccuracyBest

Multiple languages supported

Cloud Models

Maximum Accuracy — Bring your own API key

OpenAI

gpt-4o-mini-transcribe

Industry-leading accuracy with OpenAI's latest speech models.

~$0.006/minute

Groq

Whisper Large V3 Turbo

Fast, accurate results powered by Groq's inference engine.

~$0.006/minute
Multi-Language Support

Speak any language.

Dictate in 90+ languages. Switch instantly with a click or assign hotkeys for your most-used languages.

🇺🇸🇪🇸🇫🇷🇩🇪🇮🇹🇵🇹🇳🇱🇵🇱🇸🇪🇩🇰🇳🇴🇫🇮🇨🇿🇷🇴🇭🇺🇹🇷🇬🇷🇷🇺🇺🇦🇮🇱🇸🇦🇮🇳🇵🇰🇧🇩🇨🇳🇯🇵🇰🇷🇻🇳🇹🇭🇮🇩🇲🇾🇮🇷🇰🇪🇿🇦🇭🇷🇧🇬🇷🇸🇺🇸🇪🇸🇫🇷🇩🇪🇮🇹🇵🇹🇳🇱🇵🇱🇸🇪🇩🇰🇳🇴🇫🇮🇨🇿🇷🇴🇭🇺🇹🇷🇬🇷🇷🇺🇺🇦🇮🇱🇸🇦🇮🇳🇵🇰🇧🇩🇨🇳🇯🇵🇰🇷🇻🇳🇹🇭🇮🇩🇲🇾🇮🇷🇰🇪🇿🇦🇭🇷🇧🇬🇷🇸🇺🇸🇪🇸🇫🇷🇩🇪🇮🇹🇵🇹🇳🇱🇵🇱🇸🇪🇩🇰🇳🇴🇫🇮🇨🇿🇷🇴🇭🇺🇹🇷🇬🇷🇷🇺🇺🇦🇮🇱🇸🇦🇮🇳🇵🇰🇧🇩🇨🇳🇯🇵🇰🇷🇻🇳🇹🇭🇮🇩🇲🇾🇮🇷🇰🇪🇿🇦🇭🇷🇧🇬🇷🇸🇺🇸🇪🇸🇫🇷🇩🇪🇮🇹🇵🇹🇳🇱🇵🇱🇸🇪🇩🇰🇳🇴🇫🇮🇨🇿🇷🇴🇭🇺🇹🇷🇬🇷🇷🇺🇺🇦🇮🇱🇸🇦🇮🇳🇵🇰🇧🇩🇨🇳🇯🇵🇰🇷🇻🇳🇹🇭🇮🇩🇲🇾🇮🇷🇰🇪🇿🇦🇭🇷🇧🇬🇷🇸
🇸🇰🇸🇮🇪🇪🇲🇰🇱🇹🇱🇻🇦🇿🇧🇦🇰🇿🇦🇱🇮🇸🇦🇲🇳🇵🇲🇳🇺🇿🇹🇲🇹🇯🇬🇪🇧🇾🇲🇲🇰🇭🇱🇰🇿🇼🇳🇬🇸🇴🇪🇹🇫🇴🇭🇹🇦🇫🇲🇹🇱🇺🇵🇭🇲🇬🇨🇩🇳🇿🇬🇧🇻🇦🇸🇰🇸🇮🇪🇪🇲🇰🇱🇹🇱🇻🇦🇿🇧🇦🇰🇿🇦🇱🇮🇸🇦🇲🇳🇵🇲🇳🇺🇿🇹🇲🇹🇯🇬🇪🇧🇾🇲🇲🇰🇭🇱🇰🇿🇼🇳🇬🇸🇴🇪🇹🇫🇴🇭🇹🇦🇫🇲🇹🇱🇺🇵🇭🇲🇬🇨🇩🇳🇿🇬🇧🇻🇦🇸🇰🇸🇮🇪🇪🇲🇰🇱🇹🇱🇻🇦🇿🇧🇦🇰🇿🇦🇱🇮🇸🇦🇲🇳🇵🇲🇳🇺🇿🇹🇲🇹🇯🇬🇪🇧🇾🇲🇲🇰🇭🇱🇰🇿🇼🇳🇬🇸🇴🇪🇹🇫🇴🇭🇹🇦🇫🇲🇹🇱🇺🇵🇭🇲🇬🇨🇩🇳🇿🇬🇧🇻🇦🇸🇰🇸🇮🇪🇪🇲🇰🇱🇹🇱🇻🇦🇿🇧🇦🇰🇿🇦🇱🇮🇸🇦🇲🇳🇵🇲🇳🇺🇿🇹🇲🇹🇯🇬🇪🇧🇾🇲🇲🇰🇭🇱🇰🇿🇼🇳🇬🇸🇴🇪🇹🇫🇴🇭🇹🇦🇫🇲🇹🇱🇺🇵🇭🇲🇬🇨🇩🇳🇿🇬🇧🇻🇦

Quick Language Switcher

Switch languages instantly from the recording bar. Just click the flag and pick your language.

Per-Language Hotkeys

Assign a unique hotkey to each language. Press once to start dictating in that language.

Auto-Detect Mode

Let the model automatically detect which language you're speaking. Great for multilingual conversations.

One hotkey per language

Assign a dedicated keyboard shortcut to each language you use. Press your hotkey and immediately start dictating in that language — no menu navigation needed.

  • Instant language switching with a single keypress
  • Works system-wide in any application
  • Set up as many languages as you need
Language Hotkeys
🇺🇸
English
Recording...
E
🇪🇸
Spanish
Press to dictate
S
🇫🇷
French
Press to dictate
F
🇩🇪
German
Press to dictate
G
🇯🇵
Japanese
Press to dictate
J
+Add more languages in settings
Always-On Rules

Automatic text cleanup.

Toggle rules that apply to every transcription. Fix grammar, remove filler words, and polish your text automatically.

How it works

Enable the rules you want, and they'll automatically apply to every transcription. Rules combine together — enable "Fix Grammar" and "Remove Filler Words" and both will be applied. Works with AI Functions too.

Fix Grammar & Spelling

Correct grammar, spelling, and punctuation errors

Before

"i was going too the store but i didnt have no money"

After

"I was going to the store, but I didn't have any money."

Remove Filler Words

Remove 'um', 'uh', 'like', 'you know', etc.

Before

"So like, um, I was thinking that, you know, we should basically just, uh, go with option A"

After

"I was thinking that we should go with option A"

Smart Punctuation

Add proper sentence structure and punctuation

Before

"the meeting is at 3pm we need to prepare the slides also dont forget to invite sarah"

After

"The meeting is at 3pm. We need to prepare the slides. Also, don't forget to invite Sarah."

Professional Tone

Ensure a polished tone for business communication

Before

"hey can u send me that thing we talked about thanks"

After

"Hi, could you please send me the document we discussed? Thank you."

Create your own custom rules with any instructions you want.

AI Functions

Transform your voice.

Process your dictation with AI before pasting. Write emails, format code prompts, translate, summarize — all with a single click.

One-click AI processing

Select a function from the recording bar, speak naturally, and your text is automatically processed before pasting. No extra steps. Each function uses your chosen AI provider and model.

Professional Email

Clean up and professionalize email drafts

Input

"hey can we push back the meeting i have another thing that came up also did you get the docs i sent"

Output

"Hi, Would it be possible to reschedule our meeting? Something has come up that I need to attend to. Also, I wanted to confirm whether you received the documents I sent earlier. Thank you."

Vibe Coding

Transform ideas into LLM-ready coding prompts

Input

"i want to make a button that like when you click it it shows a modal and then actually no make it a dropdown instead with like three options"

Output

"Create a button component that opens a dropdown menu on click. The dropdown should display three selectable options. Style it to match the existing design system."

Summarize

Create concise summaries of your thoughts

Input

"so basically the main takeaway from the meeting was that we need to focus on the mobile experience first because most of our users are on phones and then after that we can work on the desktop version and also john mentioned we should look into the analytics to see what features people actually use"

Output

"Key takeaways: (1) Prioritize mobile experience due to majority mobile user base, (2) Desktop version comes after mobile, (3) Review analytics to identify most-used features."

Casual Rewrite

Make formal text more friendly and conversational

Input

"I am writing to inform you that your request has been processed and you should expect to receive confirmation within 3-5 business days."

Output

"Just wanted to let you know we've processed your request! You should get a confirmation in the next 3-5 business days."

Translate

Translate your speech to any language

Input

"Thank you for your help with this project. I really appreciate it."

Output

"Gracias por tu ayuda con este proyecto. Realmente lo aprecio."

Custom Functions

Create your own with any AI provider

Use OpenAI, Anthropic, Gemini, Groq, Mistral, or Grok with your own custom prompts.

Supported AI Providers

OpenAIGPT-4o, GPT-4.1
AnthropicClaude 3.5, Claude 4
GoogleGemini 2.5
GroqLlama 3.1, Mixtral
MistralLarge, Codestral
GrokGrok 3, Grok 4

RIP Subscription Wrappers

Imagine paying a monthly fee for a Whisper wrapper. Bruh.

OpenWhisper&WisprFlow

Died of recurring billing syndrome

💀

* No wrappers were harmed in the making of this landing page. Just their business models.

One-time purchase — No subscription

Pay once. Use forever.

No subscriptions. No recurring fees. No BS. One payment, lifetime access on 3 devices.

$29USD

One-time payment

3 Macs included
Lifetime updates

Because paying a subscription for a Whisper wrapper is crazy.

Everything included

  • Lifetime license — pay once, own forever
  • Use on up to 3 Macs
  • All future updates included
  • Local Whisper models (Tiny to Medium)
  • Cloud transcription (OpenAI, Groq)
  • Import & transcribe any audio file
  • Recording history with re-transcribe
  • Custom vocabulary dictionary
  • Auto-paste & auto-copy

Tired of subscriptions? Us too. That's why Sotto is a one-time purchase. No monthly fees, no annual renewals. Just pay once and it's yours forever.

Instant delivery
macOS 13+ required
License key via email
3 device activations