Why We're Moving to DeepSpeech on March 31
over 6 years ago – Mon, Jan 22, 2018 at 09:34:10 PM

Our community asked, we're answering. 

Read notes from our CTO, Steve Penrod about our move to DeepSpeech.  

From Steve;  

Misunderstandings of how Mycroft performs Speech to Text is one of the things I hear about regularly. So today I'll provide some clarity on how it works now, why it works that way, and where we are heading with this technology in the future.  

Mycroft Design Decisions  

It is rare that there is only one way to solve a problem. So every project requires thoughtful analysis of the end user's needs to select the best problem solving approach. With careful thought, experimentation and planning it is usually possible to work out a balanced approach to solve the problem. One of the key factors in choosing a solution is the design criteria. What does the user value? How does the development team best provide that value? What is the top priority? What can wait? When we started Mycroft our design criteria consisted of:  

Choice  

One of the key things we wanted to enable is choice. We knew that the technology we were using two years ago would likely not be the same technology we'd use in five years. Similarly, we knew that the technology that we decided to use would not be the technology that everyone would want to use. So we architected our system to allow choice — by both the project and by individual users — for the key technical components: Speech to Text, Text to Speech and Skills.

(read more on our blog)

Lessons Learned from Crowdfunding
over 6 years ago – Tue, Jan 16, 2018 at 10:07:26 AM

We saw it all in our first two crowdfunding campaigns. As we set to launch our second device soon (sign up for updates here!), we’ve been reflecting on the lessons we learned with the Mark I.

A history lesson (on Mycroft)

We ran our first Kickstarter in September of 2015 and finished successfully at 128% of our goal. Over the following months, we completed multiple prototypes, worked with the FCC, found some great suppliers and dealt with some bad ones. We learned a lot about international shipping and something called VAT. The result? In July 2017 we shipped out the last of the 1,500 advance prototypes.

During this entire period, our software team worked hand in hand with the Mycroft community to build open technologies for wakeword spotting, speech to text, natural language processing, and speech synthesis. We released our first alpha in August 2017 and are on track to release the software to beta in February of this year. The software is starting to get good.

Now that we’ve shipped the Mark I and the software is coming together we are confident we can ship the next generation of our technology in a much shorter time frame. We have working software, supplier relationships, a better understanding of shipping and, of course, now we know what VAT is.

We can’t thank our backers enough for helping us and staying connected with us as we navigated an extremely difficult process. What some don’t realize with crowdfunding campaigns is that you are buying into an idea, not an off-the-shelf product. Backers and creators go on the journey of bringing an ambitious idea to life together.

(continue reading on our blog)

What's New V0.9.7
over 6 years ago – Mon, Dec 11, 2017 at 02:54:58 PM

V0.9.7 Is Here.  

Thanks to our team member, forslund, for rolling this out.  

E-mail Send #1230  

The skills API now includes a self.send_email() method, allowing the skill to send an e-mail to the user. This is practical if the skill wants to send a link to the user, or for things such as shopping lists.  

Metrics #1247  

The Device API now includes a metrics method, allowing skills and core to upload information to the Mycroft server should the user allow it. The method is very general purpose and can take any method as long as it can be serialized to json.  

Skills developer can easily access this functionality using the self.report_metric method in the Skill API.  

Misc  

• Added support for Google Cloud API SST, big thanks to @ahayworth for providing this. #1226  
• The AudioService can now add files to currently playing playlist #1197  
• Listening can now be triggered from the messagebus by sending mycroft.mic.listen #1244  
• Remove pystache dependency #1235  
• Improve lip synch #1242  
• Code cleanup and bugfixes, Many thanks to @gauravMann for cleaning up the usage of mutables as default arguments.

As always, visit our Facebook Page for social updates and our blog to keep up with the latest post. If you would like to keep up with what other members are doing with Mycroft, chat with us on Mattermost, our open source chat forum.

Megabots Giant Robot Tournament
over 6 years ago – Thu, Nov 16, 2017 at 01:15:56 PM

The team that brought you the Giant Fighting Robot Duel is back with an even bigger and better project. After defeating Suidobashi Heavy Industry and winning the first ever Giant Robot Duel in history, Megabots is upping the ante by kicking off an entirely new sport.

Megabots has just launched a crowdfunding campaign to raise funds for the world's first Giant Robot Tournament. 

The team at Mycroft is incredibly excited to see the next campaign as we have been following the Megabots team since they funded their first campaign. Mycroft CEO even backed the first project and has the sticker to prove it.

You will not want to miss this campaign.

 

All Delivered
over 6 years ago – Wed, Oct 25, 2017 at 10:26:04 AM

This post is for backers only. Please visit Kickstarter.com and log in to read.