The Mycroft Mark II is the next generation open smart speaker. It sounds better, hears better, and incorporates all of the lessons learned from the Mark I.

Mark II Prototype Development Kits are not for sale. They are available for use within a Development Partnership Agreement.

Mark II smart speaker sketchesHardware Features

The Mark II has state of the art features like array microphones, beamforming, active noise cancellation, and a bright screen to display important information. It also sounds great, filling your room with vivid highs and deep lows.

Each Mark II Prototype Development Kit is hand built and subject to variation as development iterations are incorporated. Required specifications will be outlined in your Development Partnership Agreement.

Conversational Software Stack

Mycroft provides a complete conversational software stack that is both private and open.  Our tech stack includes all of the major components of a digital assistant along with a robust software and hardware development and integration testing framework.

Wake Word Spotting

The Precise Wake Word Spotter listens to the environment and ignores everything but the wake word. Our default wake word is “Hey Mycroft”, but users can customize this to fit their needs. “Hey Ezra”, “Christopher”, and “Jarvis” are all popular community wake words.


For users who opt-in, we capture commands and make them available to improve DeepSpeech by Mozilla. We will move to DeepSpeech as our primary STT engine once the speed and accuracy meets the necessary standard.

Natural Language Understanding

Our Adapt and Padatious intent parsers use both known entity rules and machine learning to determine what the user wants to do.

Natural Language Processing

Lingua Franca, our multilingual text parsing and formatting library allows Mycroft to convert natural language to common data structures. Then back to naturally expressed language for the response.


Our TTS synthesizer, Mimic, uses parametrics to synthesize voice on device. Mimic II, uses machine learning and sounds significantly more human-like than other open source libraries.

Skills Framework

The Mycroft Skills Kit along with our Voice User Interface Design Guidelines get you up and running quickly. This comprehensive toolset helps designers and developers work together to create intuitive and meaningful experiences.

Test Framework

Utilize our Voight Kampff integration test framework to enhance your development workflow, and ensure your users receive a premium experience every time.

Machine Learning

The Mycroft Community is using machine and deep learning to build an assistant that improves over time and learns how to better serve you. We’re incorporating ML into every step of the user interaction from wake word processing to speech synthesis and it’s all open source. The data being used is generated by Mycroft users who have explicitly opted in to share their data to our open data set. Unless users explicitly opt-in, we delete data as soon as the interaction is completed.

What in Voice Tech is Open Source?

Open source voice tech comparison table

Our Philosophy


Voice assistants from Big Tech know a lot about you. They know when you get home, how many people are in the house, what type of pets you have, and what your calendar looks like. Here at Mycroft, we’ve got the only voice assistant that respects your privacy. We don’t capture information about you, your family or your pets.

We don’t store voice data: Other voice assistants on the market store every voice snippet they hear. At Mycroft, we delete the recordings as they come in, unless you actively choose to share your data by opting into our open data set.

We don’t create targeted ads based on your interactions: The difference between us and the Big Tech Giants is our business model. We want to provide voice services to our members, the others have business plans around targeting you with advertisements or selling you goods and services.

We don’t create ads or sell your data to anyone who does. Our software is open so anyone can inspect the code and see exactly what we are doing with user data. This transparency ensures good behavior and creates trust between Mycroft the company and Mycroft the community.


Create a voice assistant that suits you. From its name to its voice, to its personality, we are creating tools that allow you to create an agent that represents you.

Personality: An anonymous black cylinder no more; Mycroft has a relatable, easy to interact with personality. To customize your Mark II, you can choose different home screen options such as weather and time widgets.

Custom voices: You can choose the voice of your Mycroft, currently we support American Male, British Male, and a range of third-party TTS services.

Custom wake words: You can further customize your Mark II by training a unique wake word.

Open Source Wins

The Switzerland of voice assistants, Mycroft is a neutral player in the tech ecosystem wars. We’re not connected to any services. The recent media streaming/e-commerce feud between Amazon and Google left consumers out in the cold. With Mycroft, you never have to worry about the giants fighting over rights, we are a neutral player in the game allowing you to use what you want. We’re the Switzerland of voice assistants, allowing users to choose the best application for their need; we have no strategic reason to funnel users into any brand’s service.

Open code for anyone to contribute. You already know that being open source makes Mycroft customizable and transparent, but did you know open software breeds diversity? Building an AI that anyone can contribute to means making technology that is more representative of the world we live in today, a technology that is equitable, neutral, forward thinking.

Some of our partners and supporters. We’re humbled by their support.

Mycroft has been featured in a diverse array of publications.