Murmure

A privacy-first, open-source speech-to-text application that runs entirely on your machine, powered by a neural network via NVIDIA’s Parakeet model for fast, local transcription. Murmure turns your voice into text with no internet connection and zero data collection, and supports 25 European languages.

Learn more on the on the official website.

Features

• Privacy First: All processing happens locally on your device. No data ever leaves your computer.
• No Telemetry: Zero tracking, zero analytics. Your data stays yours, always.
• Open Source: Free and open source software. Inspect, modify, and contribute.
• Powered by Parakeet: NVIDIA’s state-of-the-art speech recognition model runs entirely on-device for fast, low-latency transcription.

Supported Languages:

Bulgarian (bg), Croatian (hr), Czech (cs), Danish (da), Dutch (nl), English (en), Estonian (et), Finnish (fi), French (fr), German (de), Greek (el), Hungarian (hu), Italian (it), Latvian (lv), Lithuanian (lt), Maltese (mt), Polish (pl), Portuguese (pt), Romanian (ro), Slovak (sk), Slovenian (sl), Spanish (es), Swedish (sv), Russian (ru), Ukrainian (uk)

Installation

Windows (Official)

⚠️ Windows SmartScreen : This installer is not signed with a commercial certificate (which costs ~€200–€500/year).
If you downloaded it from our official GitHub releases, you can safely continue.

🛡️ We guarantee the installer is safe, contains no malware, and you can verify the source code or even compile it yourself if you prefer.

1. Download Murmure_x64.msi from the release page
2. Run the installer and follow the setup wizard.

Linux (Official)

⚠️ Murmure currently has limited support on Wayland-based distributions (except Fedora, which can fall back to X11 for some apps).
This appears to be related to Wayland’s sandbox restrictions for AppImages, the global shortcut to start recording will not work in this environment.
No workaround is available yet. See #28

1. Download Murmure_amd64.AppImage from release page
2. Make it executable: chmod +x Murmure_amd64.AppImage
3. Run the AppImage.

Murmure uses the ALSA API to access your microphone, so if you're running Pipewire for your audio stack, make sure that the ALSA API calls are routed through it (e.g. by installing the pipewire-alsa package on Arch Linux), otherwise you'll have errors such as ALSA lib pcm_dsnoop.c:567:(snd_pcm_dsnoop_open) unable to open slave.

MacOS (Official)

⚠️ MacOS may show security warnings because Murmure isn’t signed with a paid Apple certificate. These warnings are expected for independent apps, and Murmure is safe to install.

🛡️ We guarantee the installer is safe, contains no malware, and you can verify the source code or even compile it yourself if you prefer.

1. Download Murmure_aarch64_darwin.dmg from the release page
2. Open the DMG. If macOS blocks it, go to System Settings → Privacy & Security and click "Open Anyway".
3. Drag Murmure to the Applications folder, then open it from there.
4. If you see an "app is damaged" message, click Cancel, run xattr -cr /Applications/Murmure.app in Terminal, then reopen Murmure.

Arch Linux (Community)

⚠️ Community builds are maintained by kind contributors on a best-effort basis. They do their best to keep them up to date, but there’s no guarantee they will always be. If you encounter a bug with one of these packages, please open an issue in the corresponding community repository instead.

Community repository: https://github.com/Horgix/aur-package_murmure_mirror

On Arch Linux, you can install the murmure package directly from the AUR using your favorite helper:

aura -A murmure
# Or
yay -S murmure
# Or
paru -S murmure

Usage

Murmure provides a clean and focused speech-to-text experience. Once launched, simply start recording your voice. The text appears instantly, processed directly on your computer.

Typical use cases include:

• Dictating to any AI prompt (Cursor, ChatGPT, Mistral, etc.)
• Writing notes hands-free
• Capturing creative ideas or dictation

Because all computation is local, no network connection is required.

Technology

Murmure uses NVIDIA’s Parakeet TDT, a highly optimized, experimental transformer-based speech recognition model designed for low-latency, on-device inference. It combines fast transcription with strong accuracy across multiple languages, running efficiently on consumer GPUs or CPUs.

Changelog

See CHANGELOG.md.

🗺️ Roadmap

• (1.6.0) Major: Connect a local LLM (ollama) to enhance or modify transcriptions (post-processing)
• (1.6.0) feat(ui): Improve UI of notifications
• (1.6.0) feat(overlay): Add a song option when start & stop recording
• (1.6.0) feat: Profile preset (general, medical, developper, translator) that prefill dictionary and LLM instructions
• (1.6.0) feat(shortcuts): Add keyboard shortcuts to start and stop recording (outside push-to-talk mode)
• (1.6.0) refactor: folder structure & split audio.rs
• (1.6.0) refactor(dictionary): split dictionary into dictionary.json
• (1.6.0) feat(dictionary): Make dictionary visible for LLM Connect via {{DICTIONARY}}
• (1.6.0) feat(rules): Format & Actions dictionary to choose how to add line breaks, trailing-space, convert text-to-number or fill content when saying specific trigger words
• feat(dictionary): import/export words from dictionary (medical preset and other)
• poc: Portable version without installer (& admin password) for hospital
• refactor(dictionary): Improve the custom dictionary algorithm (performance and algo)
• feat(shortcut): Add a shortcut to add automatically a word in dictionnary after selecting it (copy selection > read word > add it to dictionary)
• feat(llm): Allow multi LLM setup width different shortcut
• fix(overlay): prevent launching multiple Murmure App instances when clicked rapidly
• fix(overlay): weird position on some screen Kieirra#64
• feat(overlay): be able to drag-n-drop the overlay to change position Kieirra#64
• feat(overlay): Add a size option
• feat : be able to pin Murmure to the dock Kieirra#64
• fix(visualizer): dynamic or lower sensibility
• fix(visualizer): the visualizer does not always reset at the end of a transcription
• feat(settings): Allow selecting the input microphone
• refactor: Secure Settings Persistence (Migrate to tauri-plugin-store for atomic writes)
• fix(shortcuts): Improve available shortcuts on Linux & Windows
• feat: add .deb file and register it on debian/ubuntu/"linux mint" package manager
• (under consideration) feat(webhook): Send an HTTP request after CTRL + SPACE, opens up many interesting possibilities
• Major: Implement a plugin system
• Major(under consideration): Drag & drop support for audio files (MP3, WAV) with automatic transcription (and maybe speaker diarization)
• Major(under consideration): Add real-time streaming (POC)

Acknowledgments

• Thanks to NVIDIA for the Parakeet TDT model, Tauri for being an amazing tool, and to the open‑source community for their tools and libraries.

License

Murmure is free and open source, released under the GNU GPL v3 License. You can inspect, modify, and redistribute it freely as long as derivative works remain open source.

Contributing

See CONTRIBUTING.md.

Reporting issues is done on GitHub.

Support Development

If you like Murmure and want to support its development: Support on Tipeee