Spoke

Key Features

• •On-Device Transcription: Spoke uses a 600M parameter neural network that runs entirely on your Mac, ensuring your voice never leaves your computer. Transcription is sub-400ms fast and works offline, offering both privacy and speed.
• •Flow Builder: Create custom multi-step workflows by chaining voice input with actions like AI skills, app switching, text transforms, and webhooks. Each flow can be triggered by its own shortcut, enabling powerful automation from a single voice command.
• •AI Post-Processing: Attach AI Skills to your flows to fix grammar, format context, translate languages, or turn speech into code prompts. Supports OpenAI, Anthropic, Gemini, and Ollama with your own API key, and you can switch providers freely.
• •Universal App Compatibility: Spoke works in every macOS app without any integrations—just hold a key and speak. Text appears at the cursor in Notion, Slack, Terminal, VS Code, Xcode, and any other application.
• •Coding Agent Integration: Speak, screenshot, and copy into Claude Code in one message via a direct encrypted pipeline. Timeline-aware screenshots and clipboard capture provide rich context for AI coding assistants.
• •24 Language Support: Dictate in 24 languages with full on-device privacy. Switch languages without any configuration or internet dependency, making Spoke suitable for multilingual users.
• •Menu Bar Access: Spoke lives in your Mac’s menu bar, always ready with a lightweight footprint (<50MB RAM, 1-2% CPU at idle). It is a native macOS app built with Apple frameworks, not a web wrapper.

Use Cases

• •Email and Professional Communication: Dictate a draft email or reply casually, then use Spoke’s AI skill to fix grammar and adjust tone before pasting into your mail client. This saves time and ensures polished output.
• •Coding and Development: Describe a function or bug verbally while Spoke transcribes it directly into your IDE. Use the Flow Builder to format speech as CLI commands and execute them in Terminal, all hands-free.
• •Content Creation and Blogging: Dictate first drafts of blog posts, articles, or social media updates at three times the speed of typing. Spoke inserts text into any editor, allowing you to capture ideas without interrupting your flow.
• •Team Communication and Project Management: Speak messages into Slack, Discord, or Notion without switching windows. Set up per-app flows to automatically adjust tone or add formatting, keeping communication fast and consistent.
• •Talk to AI Coding Agents: While developing, hold a shortcut to speak, take a screenshot, and copy an error—all in one voice command. Spoke sends the full context (timestamped screenshots, clipboard) to Claude Code via an encrypted channel for richer AI interactions.
• •System Automation and Workflow: Create a flow that captures a voice command, switches to a specific app (e.g., Calendar), types a preformatted message, and presses Return. This automates repetitive tasks like booking meetings or logging time.

Who is this for?

Spoke is designed for macOS users who type extensively and want to speed up their workflow with voice. Primary segments include software developers who need to write code, documentation, and team messages without leaving their keyboard; writers and bloggers who draft long-form content quickly; project managers and remote workers who send frequent Slack messages and emails; and anyone who uses multiple apps throughout the day and values privacy. It also targets users who communicate with AI coding agents (like Claude Code) and need a fast, contextual voice input pipeline. Spoke appeals to privacy-conscious individuals who avoid cloud-based dictation services and prefer a one-time purchase model over ongoing subscriptions.