Verified project record

OpenWhispr/openwhispr

OpenWhispr is a free, open-source Electron desktop app for voice-to-text dictation, meeting transcription, and notes, positioned as an alternative to WisprFlow and Granola without vendor lock-in. It offers local offline transcription via OpenAI Whisper and NVIDIA Parakeet models, with optional cloud processing and sync that requires your own API keys.

Project overview

It combines global-hotkey dictation into any app with local offline transcription and an MCP server for connecting AI assistants to your notes, addressing privacy concerns around commercial dictation tools.

Project type: AI Agent · Audio & Speech
Use cases: Meeting Notes · Audio & Speech
Deployment: Refer to project documentation
License: MIT

Best for

General users with a desktop environment on macOS, Windows, or Linux who want free, open-source voice dictation, transcription, and notes without vendor lock-in or data collection, and who can run transcription locally with optional GPU acceleration.
Users comfortable supplying their own API keys when they choose to enable optional cloud models and sync features.

Key capabilities

Press a global hotkey, speak, and the dictated text is pasted into any app at the cursor; a dedicated translation hotkey is also included.

Limitations and risks

Cloud sync and AI features rely on external providers, and BYOK (bring your own key) API keys are needed for cloud models. Users who enable these features send data to those providers.
As an Electron desktop app, it is not suitable for headless or non-desktop usage.
Cost dependency and telemetry practices are not documented, and minimum hardware requirements are not specified. GPU support is optional.

Getting started

Download the prebuilt installer from the releases page (no build steps required), install and launch the app, then press the hotkey and dictate. A community-maintained Homebrew cask is also available for macOS and x86_64 Linux.

Alternatives and comparisons

A local-first application for audio and video transcription, structured Markdown note generation, and content-based Q&A, useful if you need file-based transcription rather than system-wide hotkey dictation.
Local-first meeting capture with on-device transcription into user-owned Markdown files, searchable via Claude Code, Codex, Cursor, or any MCP client.
An on-device toolkit for real-time voice tasks including speech-to-text, text-to-speech, voice cloning, speaker identification, and intent recognition, requiring no external API keys or accounts.

Project comparisons

Evidence and sources

README: Privacy-first voice-to-text dictation with AI agents, meeting transcription, and notes. Cross-platform for macOS, Windows, and Linux.
README: | macOS (Apple Silicon) | [`.dmg`](https://github.com/OpenWhispr/openwhispr/releases/latest) |
README: - **Voice dictation** — global hotkey to dictate into any app with automatic pasting
README: git clone https://github.com/OpenWhispr/openwhispr.git cd openwhispr npm install npm run dev
README: **Public API & MCP** — manage notes and transcriptions programmatically or connect your AI assistant via the [MCP server](https://docs.openwhispr.com/integrations/mcp)

Source updated at: 2026-10-08T18:20:55Z

Open original GitHub project