mylxsw/typeflux

Typeflux is a macOS menu bar voice input tool built with Swift. It is designed for a fast "hold to talk, release to insert" workflow: press a hotkey, speak naturally, let the app transcribe your speech, and send the resulting text back into the currently focused app.

What it solves

Typeflux is a macOS application that eliminates the friction of voice-to-text by injecting transcribed text directly into any active application at the cursor position. It removes the need to switch between a dictation app and the target application, allowing users to input text roughly four times faster than typing.

How it works

Users interact with the app via hotkeys: holding the Fn key allows for natural speech dictation, while double-pressing Fn activates an AI assistant for Q&A, rewriting, and translation. The app supports both cloud-based speech-to-text (STT) providers (such as OpenAI Whisper, Groq, and Google) and fully local on-device models (such as SenseVoice, WhisperKit, and Qwen3-ASR) to ensure privacy and offline functionality.

Who it’s for

MacOS users who want a fast, seamless voice input method for any text field in browsers, code editors, and terminals, as well as those who prioritize privacy by running models locally on their machine.

Highlights

  • Direct Text Injection: Transcribes and inserts text immediately into the active app without copying and pasting.
  • Local-First Privacy: Supports fully on-device inference for voice and text processing, meaning data never leaves the Mac.
  • AI Agent Capabilities: Beyond simple dictation, it can rewrite selected text or handle complex workflows through voice commands.
  • Multiple Backend Options: Offers a wide range of local and cloud providers for varying accuracy and latency needs.
  • Streaming Preview: Provides real-time feedback by showing partial transcriptions while the user is speaking.

Related

  • Project
    altic-dev/FluidVoice

    An open-source voice-to-text dictation app for macOS that enables on-device AI transcription and system control via voice commands.

  • Project
    moinulmoin/voicetypr

    An offline-first voice-to-text dictation app for macOS and Windows that transcribes speech and inserts it directly into any active application.

  • Project
    TypeWhisper/typewhisper-mac

    TypeWhisper is a macOS app that transcribes speech to text using on‑device or cloud models, lets you run AI‑powered workflows (translation, rewriting, app automation), and offers extensive customisation via plugins, a local HTTP API, and a CLI. It supports system‑wide dictation hotkeys, calendar‑aware meeting automation, dictionary/snippet personalization, and premium features like cloud sync and correction learning.

  • Project
    watzon/pindrop

    An open-source, Mac-native AI dictation app that converts speech to text using local Apple Silicon engines or optional cloud services for maximum privacy.

  • Project
    OpenWhispr/openwhispr

    OpenWhispr is a cross‑platform desktop app that converts speech to text, provides live meeting transcription with speaker diarisation, and lets you talk to AI assistants (GPT‑5, Claude, Gemini, etc.) directly from your keyboard. It works offline with Whisper‑based models or via cloud services, stores notes locally (with optional cloud sync), and offers an API for programmatic use.