WhisperScribe ProLocal transcription, private

SonaVoice

SonaVoice is a Windows voice-typing tool: place your cursor in any app, hold Right Ctrl to speak, and it inserts clean, punctuated text, with a custom dictionary for names and jargon.

uho Dictation

uho Dictation is a macOS dictation app that turns speech into text inside any application. It runs OpenAI Whisper models locally on Apple Silicon, so audio never leaves the device and the tool works without an internet connection or account. A single Fn keypress starts and stops recording, and the transcribed text is inserted directly at the cursor in the active app. It targets Mac users who dictate across many apps through the day, and is sold as a one-time lifetime license instead of a monthly subscription.

Synopsule

Synopsule is a 1.99 USD one-time Mac and iPhone app that records meetings and runs Whisper on device to produce searchable, speaker-labeled transcripts.

VidTranscriber

VidTranscriber transcribes YouTube links, uploaded video and audio files, and Zoom recordings into searchable text with speaker labels, AI chapters, and exports in TXT, SRT, VTT, Markdown, and DOCX.

Steno: Open-Source AI Note-Taking for High-Confidentiality Conversations

Steno is an open-source AI note-taking tool designed for high-confidentiality conversations. It runs locally or on private servers, automatically generating structured notes from meeting recordings while ensuring data never leaves your controlled environment. Ideal for government, defense, legal, and executive teams. Primary language is TypeScript, license is MIT.

vexa: Open-source meeting transcription API

vexa is an open-source meeting transcription API that seamlessly integrates with Google Meet, Microsoft Teams, and Zoom. It automatically joins meetings and provides real-time transcription via WebSocket. An MCP server enables AI agent integration. Users can self-host or use the SaaS offering. Written in Python, vexa has over 2500 stars on GitHub, making it a robust solution for automating meeting documentation and AI-powered analysis.

typewhisper-mac: on-device speech-to-text for macOS

typewhisper-mac is an open-source macOS app for on-device speech-to-text. It leverages local AI on Apple Silicon or Intel for real-time, privacy-focused transcription without an internet connection. Supports multiple languages including Chinese and English, with an optional cloud mode for enhanced accuracy. Built with Swift and licensed under GPL-3.0, it has over 1500 GitHub stars.

Handy: Turn keyboard shortcut into local speech-to-text

Handy is a cross-platform desktop app that turns a keyboard shortcut into speech-to-text: press, speak, and your words land in whatever text field has focus. Everything runs locally with no cloud upload, using Whisper or Parakeet V3 models and Silero voice activity detection. Built with Rust and Tauri on the backend and React with Tailwind on the front end, it ships for macOS, Windows, and Linux under the MIT license. As of collection time, it had 23,022 GitHub stars.

amical: Local-first AI dictation for offline speech-to-text

amical is an open-source, local-first AI dictation application that leverages open models like Whisper for fast, accurate, and offline speech-to-text. It claims to triple typing speed without a keyboard, supports multiple languages, and prioritizes user privacy by processing all data on-device. Built with TypeScript, it is suitable for both developers and general users.