Features
Sotto features
Dictation is the core. Sotto also runs a voice assistant, a transcription server for your other machines, and whichever models you trust.
$49 once, up to 3 Macs, all future updates included. See pricing
Dictation
System-wide voice typing that runs on your Mac.
Push-to-talk or toggle
Hold a hotkey to talk, or tap to start and stop.
Types into any Mac app
The transcript lands where your cursor is. Auto-copy is there if you prefer the clipboard.
Local by default
Whisper and Parakeet models run on your Mac. After a one-time model download, dictation works offline.
Long dictations, streamed
Sotto transcribes while you talk, so a long take does not wait until the end.
90+ languages
Switch languages from the recording bar, assign a hotkey per language, or let the model auto-detect what you are speaking.
Dictionary and find & replace
Teach Sotto your names, products and jargon, and add find-and-replace rules, regex included, for terms and formatting.
Always-on AI rules and functions
Fix grammar, remove filler words, set punctuation and tone, or run a one-click email, summary, translation or custom prompt.
Quiets your Mac, survives crashes
Other audio is muted while you record, and a recovery journal protects recordings from crashes and hangs.
Sotto Assistant
Hold a key and ask. It acts through native Mac tools, sees your screen when you allow it, and answers out loud.
77 built-in Mac tools
Calendar, Reminders, Mail, Messages, Notes, Contacts, Maps, Weather, web, image and YouTube search, the browser, Finder, apps and windows, the clipboard, selected text, screenshots, Shortcuts, media and volume, alarms and timers, and policy-gated bash.
Screen context
Optionally attach a screenshot to your request, so you can ask about whatever is in front of you.
Talks back
Replies out loud with local voices, OpenAI or ElevenLabs.
Your own MCP servers
Sotto is an MCP client: add the MCP servers you already run and their tools sit next to the built-in ones.
Guardrails built in
Bash runs behind a policy, every tool has privacy defaults, and the assistant keeps a history you can review.
Local transcription server
A local REST API so your scripts, other Macs and anything that can make an HTTP request use the same on-device models.
37 endpoints
Start and stop recording, transcribe and upload files, read history, manage AI functions, rules, dictionary and find-and-replace, speak text, and list voices, models and languages.
OpenAPI spec and Swagger docs
Browse and try every endpoint at /docs.
A token per device
Bearer-token auth with a separate token for each device, plus switches that turn whole endpoint groups off.
Wi-Fi or Tailscale
Allow remote connections and reach Sotto over your LAN or Tailscale. The pairing QR code carries both addresses, and Tailscale is preferred.
The sotto CLI
Status, models, AI functions, languages, voices, speak and transcribe from any shell. Add --host to use another Mac.
On your Mac
The API server runs on your Mac on port 8765. Remote access is opt-in.
Models and providers
On-device transcription, cloud providers with your own key, or the coding agent you already pay for.
On-device transcription
WhisperKit models from Tiny through Large V3 Turbo, Distil Large V3, and NVIDIA Parakeet v2/v3 run on the Neural Engine.
Local AI without a key
Ollama, LM Studio and Apple Intelligence run AI locally.
7 cloud transcription providers
OpenAI, Groq, xAI Grok, Deepgram Nova-3, AssemblyAI, Cartesia Ink or any OpenAI-compatible endpoint, with your own key.
15 AI providers
Including OpenAI, OpenRouter, Anthropic, Gemini, Groq, Mistral and xAI, with your own key.
Your coding agent as the backend
Claude Code, Codex, OpenCode, Grok CLI or Cursor Agent, with no API key.
Switch models anytime
Re-transcribe any saved recording with a different model, and point AI rules and functions at another provider.
Meetings and automation
Capture system audio, auto-transcribe watched folders, drive Sotto from Shortcuts, and have it read text aloud.
Meetings, calls and videos
Record system audio, or run continuous mic and system recording side by side. It happens on your Mac, so no bot joins the call.
Imports and watch folders
Drag in audio or video, or drop it into a watched folder: Sotto transcribes it, writes a title and a summary, and notifies you when it is done.
Shortcuts actions
Start, stop or toggle recording, record in a language, copy the latest transcript, transcribe the clipboard or a selected file, run an AI function, speak text, and start or stop the API server.
Text-to-speech
Read text aloud with local voices, OpenAI or ElevenLabs, or speak to a file through the API.
Modes
Named profiles, each with its own icon, colour, local transforms and AI function.
History with re-transcribe
Replay, search and re-transcribe every recording, and keep track of what cloud transcription cost.
License and requirements
One purchase covers every feature on this page.
One-time $49 purchase
No subscription and no annual renewal.
Up to 3 Macs
One license covers up to 3 Macs.
All future updates included
Updates ship with the same license.
macOS 15 Sequoia or later
Sotto is a native Swift app. Apple Silicon Macs get the best performance because local models run on the Neural Engine.
License key by email
You get a license key by email right after checkout.