Community-Driven Development

Open Voice OS is powered by developers from Linux and FOSS communities worldwide, creating an open platform for cutting-edge voice assistance technology.

Benefit from a collaborative ecosystem that continuously improves the platform.

Runs on Diverse Hardware

Supports Raspberry Pi models (3B, 3B+, 4, Zero 2), Mark I and Mark II devices, Linux desktops/laptops, and Docker containers.

Deploy voice AI on embedded devices, single-board computers, or your existing Linux system.

Offline-Capable Core

ovos-core does not require a backend internet server by default. Some skills work entirely offline, and self-hosted STT options like Vosk or Mozilla DeepSpeech are available.

Maintain control over your data with optional offline operation.

Plugin-Based STT and TTS

Choose from multiple speech recognition and text-to-speech engines, including local options like Mimic3 and online services like Google TTS or Amazon Polly.

Customize voice interactions to match your privacy and quality needs.

Skills and Customization

Extend functionality with community-developed skills from the marketplace, create your own skills, and customize wake words, UI, and more.

Tailor the assistant to your specific use cases and devices.

Free and Apache 2.0 Licensed

Open Voice OS is completely free and open source, allowing personal or commercial use, modification, and distribution.

Use, modify, and contribute without licensing costs.

• Community-Driven and Open SourceOpenVoiceOS is a community-driven, open-source voice AI platform. It is released under the Apache License 2.0, allowing free use for personal or commercial purposes, modification, distribution, and contribution back to the project.
• Multi-Platform SupportOpen Voice OS can run on a variety of hardware platforms, including Raspberry Pi models (3B, 3B+, 4, Zero 2), Mark I and Mark II devices, Linux desktops and laptops, and Docker containers. It also offers a prebuilt image for embedded systems and low-spec headless devices.
• Flexible Installation MethodsOVOS offers two main installation methods: Docker for a stable and easy setup, and Python virtual environment for developers who want to modify the code. A simple one-line installer command is provided for Docker installation.
• Plugin-Based STT and TTS ArchitectureOpen Voice OS uses a plugin-based architecture for speech recognition (STT) and text-to-speech (TTS). Default STT uses online services, but local/offline options like Vosk or Mozilla DeepSpeech are available. TTS options include Mimic3 (local), Google TTS, Amazon Polly, and more.

• Offline Functionality LimitationsWhile ovos-core does not require a backend internet server by default, the default speech-to-text (STT) engine currently requires an internet connection. Some skills can be accessed entirely offline via command line, but individual skills and plugins may require internet for a full experience.
• Relationship with MycroftOpen Voice OS and Mycroft are fully independent projects. OVOS started as MycroftOS and maintains an enhanced reference fork of Mycroft-core. It is described as the unsanctioned 'Mycroft Community Edition', and while many changes are submitted to Mycroft, they are included at Mycroft's discretion.
• Hardware Compatibility SpecificsSupported hardware includes specific Raspberry Pi models (3B, 3B+, 4, Zero 2), Mark I and Mark II devices, Linux desktops/laptops, and Docker containers. For Mark II, a specific image must be downloaded. Always check the downloads page for specific images and installation instructions.
• Grant and FundingOpenVoiceOS has received an NGI Zero Commons fund grant. Donations are appreciated but entirely optional. The project is free to use, but no specific pricing or commercial offerings are mentioned on the website.

DIY Smart Speaker

Build a do-it-yourself smart speaker using Open Voice OS on embedded headless devices or single-board computers with screens. The platform supports Raspberry Pi models (3B, 3B+, 4, Zero 2) and offers prebuilt images for low-spec hardware.

Offline Voice Assistant

Run a voice assistant without a backend internet server. While the default STT engine requires internet, you can configure local/offline options like Vosk or Mozilla DeepSpeech, and use Mimic3 for local text-to-speech.

Smart Home Control

Control smart home devices using voice commands. Open Voice OS supports customizable wake words and can be extended with community-developed skills from the marketplace.

Experimental Voice AI Development

Use Open Voice OS as a platform to test experimental features before they land in upstream Linux-based voice assistant projects. It is open to developers and contributors wanting to support specific devices or platforms.

Sign up for a free account and upload a sample of your desired brand voice.

Use the Voice Studio to fine‑tune pitch, timbre, and emotion settings.

Generate an API key and integrate the endpoint into your call platform or CMS.

Test a live call or batch job, then monitor results on the dashboard.