freestyle-voice/freestyle

What it solves

Freestyle is designed to reduce the friction of managing tasks and writing. It solves the problem of manually hunting for updates in various apps (like GitHub issues or invoices) and the clunkiness of traditional voice-to-text tools by providing a proactive, AI-powered desktop application that handles both monitoring and polished dictation.

How it works

The app integrates with a wide range of services via connected apps and the Model Context Protocol (MCP), allowing it to check for updates on a schedule and notify the user when attention is needed. For writing, it uses a voice dictation system with a hotkey-driven workflow where raw speech is processed through AI cleanup, custom prompts, and dictionary replacements to produce polished text that can be pasted directly into any application.

Who it’s for

Professionals and power users who want to integrate their voice, writing, and task monitoring across multiple desktop applications without leaving their current workflow.

Highlights

• Proactive Monitoring: Scheduled tasks that check for updates in connected apps and notify the user when action is required.
• Connected Ecosystem: Integration with Gmail, Slack, GitHub, Notion, and other MCP servers, with credentials stored locally.
• Remix Workspace: A dedicated agent workspace for research, drafting, and writing based on selected text.
• Polished Dictation: High-quality voice-to-text with translation, AI cleanup, and custom formatting options.
• Human-in-the-loop: Explicit approval is required before the AI takes any action, such as sending emails or running commands.
• Extensibility: Support for plugins and a dedicated SDK to extend the transcription pipeline.