Overview

Spoke is a native macOS dictation app that transcribes your voice into any text field using on-device processing. Hold a keyboard shortcut, speak, and the text appears wherever your cursor is — no audio ever leaves your Mac. Powered by the Parakeet speech model running locally, Spoke supports 24 transcription languages with auto-detection. Optional AI enhancement lets you plug in your own API keys for OpenAI, Anthropic, or Google Gemini to process speech on the fly. Features include Voice ID to train recognition of your voice, Flows for building multi-step voice workflows, and auto-return for instant message sending in chat apps.

Architecture: Apple Silicon

Key Features

On-device speech recognition using Parakeet model — no audio leaves your Mac
Hold a key, speak, release — text appears at your cursor in any app
24 transcription languages with auto-detection of system language
Voice ID trains Spoke to recognize your voice for hands-free mode
Flows: build multi-step voice pipelines with AI skills, text transforms, AppleScript, Shortcuts, and webhooks
Auto-return mode sends messages instantly in chat apps like a walkie-talkie
BYOK support for OpenAI, Anthropic (Claude), and Google Gemini APIs
Works in any text field — Slack, iMessage, terminal, browsers, and more
No accounts, no cloud, fully private by design