How a Self-Taught Programmer Came to Create Mugsy, an Automatic Coffee Maker

mugsy coffee maker

Matt Oswald, creator of Argyle Labs, has been on a mission to create Mugsy, a fully open-source and “hackable” pour-over coffee maker. It can learn how to make coffee from its users as they dial in the settings manually, allowing people who just want a quick cup of coffee to be left with a very satisfying brew. Not only does it make a great cup of coffee, it allows you to command it through Slack Messenger, Twitter, and whatever else you can think of!

What is Argyle Labs? Did it come before birthing the idea of Mugsy?

Argyle Labs is a company that develops open source hardware in the consumer electronics space.  One of the major aims of the company is to recreate things we use every day without all the baggage that typically comes with owning them. Whether that’s cost, functionality, usability or in some cases even things like DRM protections. Mugsy was actually the first hardware project I started working on about 3 years ago and predates the creation of Argyle Labs. At that point it was just another R and D resource sink at my software company, Argyle CO. I broke off the hardware design work into Argyle Labs a little over a year ago as it began to take over the majority of my time and I had a better understanding of what I wanted to accomplish in that space.

A post shared by M. Argyle (@marrrgyle) on

When did you first start tinkering with around with code and hardware? Did one come before the other, or did they just sort of converge over time?

I’ve been writing code since I was a little kid. Computers were around the house from an early age. This was the mid-80s so in my house it was systems like the Apple IIe and the Commodore 64 and eventually the IBM clones that started coming out. As I got older there were lots of diversions and new passions but I always came back to development, as it was the only thing that scratched the technology itch while also being wholly creative. Hardware was still a black box for me for many years. I was really interested in it but it seemed like magic. I just didn’t get it. But as the internet started to really grow, I discovered all these brilliant people sharing their work and basically showing off how the trick was done. It was enough info for me to grab and hold on to and slowly build up a deeper understanding. I’m completely self-taught and the biggest hurdle when that is how you learn is figuring out what you don’t know. Once you sort that out you can break your problem down and start grinding away at it. After a couple of years of running a team at Amazon and doing the consultant dev thing I was completely burned out on software. I needed my work to connect with the physical world in a way that software couldn’t do on its own. I was in full on existential crisis mode! At that point I made a decision to only work on things I could physically hold in my hand. I’m not 100% there yet as the majority of my income still comes from my software business but I’m a lot closer than I was yesterday.

What’s your favorite coffee brand? Do you prefer going to a café, or making it yourself?

Tough question. My favorite coffee and cafe is Cafe Ladro on 15th Ave in Seattle. Now that I’m back living in NYC, I haven’t found my home coffee spot so I mostly make my coffee at home. Lately I’ve been drinking a ton of Counter Culture Coffee’s Fast Forward. Perfect for a morning pour over, not too heavy and super tasty.  I’m also not the type of person to sit in a cafe, I get my cup and I go home and work.

A post shared by M. Argyle (@marrrgyle) on

Your Instagram mentions you like to ‘make bullsh*t and punk rock’, are you interested in creating music? What are some other hobbies of yours besides code and hardware?

Yes, music is definitely a huge part of my life. I grew up playing in punk rock bands during a really wonderful period of time and was lucky to be a very small part of an incredible music scene. The lessons I learned through my involvement in independent music informs all of the work I do today. The idea that do it yourself is not simply a necessity but more importantly a freedom. That DIY never has to mean “not as good” and can go places the giants can’t. I still play and write every day and go to tons of shows but it’s much more of a personal pursuit now instead of trying to push a band forward and tour. Other than that, I try to read as often as I can while building relationships with people who have pools I can eventually swim in. I’m really into art, astronomy, museums, local NYC history and sending pics of cute cats to my girlfriend.

Why design a coffee maker, what’s wrong with the ones on the market?

For me the price-performance ratio is just way too high to get a decent fresh cup of coffee without making it yourself. You need to grind right before you brew or you’ve already lost the war. Lower end coffee machines with grinders typically have a blade grinder when you really need a burr grinder. The temperature control is typically non-existent and the water flow control is not even considered.

So you have these very expensive machines that really don’t do much more than a low-end drip coffee maker. I’ve taken apart tons of coffee machines and the parts are mostly the same across any sub $1500 non- espresso machines. So, what are we paying for? It’s certainly not usability or options.

The plan of Mugsy is to both quickly brew a cup of coffee, but also to learn from other users about ‘how’ to make a nice cup of joe. Can you elaborate on that functionality a little bit more?

Basically, I look at Mugsy having two types of users. Those that obsess over coffee and will dial in every setting – from grind size, to bloom time and pour over pattern, and those that just want a fast cup of coffee that tastes good. The work the aficionados put in is shared with all other Mugsy users through an extensive coffee and brewing database. That combined with knowing what beans are being used allows Mugsy to always have really incredible default brewing settings. You can scan your bag of beans and Mugsy will recognize that type of coffee. If it is an unknown type a user will be able to enter basic settings like region, blend, etc to lock on to similar recipes. Once that’s done that type of coffee is now available for every user.  The goal is to minimize friction so it all happens behind the scenes. I don’t want someone who is late for work to have to hunt and peck to get what they want.

A post shared by M. Argyle (@marrrgyle) on

There is also some interesting machine learning being used to dial a cup in for specific user’s tastes. This happens by logging many environmental variables and varying them from cup to cup while asking a user if yesterday’s cup was better than todays. This functionality is optional and Mugsy is quite happy to keep quiet and stick to the basics if that’s what a user prefers.

One of the major features of Mugsy that has resonated with me is that I can create coffee without having to get out of my chair. Other than Twitter and SMS, how do you plan on interfacing the user with the Mugsy wirelessly?

Mugsy has a really simple open-ended API so it’s easy to hook into lots of different communication methods. I’ve written a Slack integration that is really nice for the workplace. There’s an Alexa skill so you can brew totally hands free from anywhere. It was important for me to be able to utilize the tech people already use to communicate to speak to Mugsy. I didn’t want to lock users into an app as that creates its own frictions and usability issues. I imagine this list will grow pretty quickly as it’s the most accessible thing for a new user to hack on their own.

A post shared by M. Argyle (@marrrgyle) on

While you have mentioned that Mugsy is currently only for pour-over coffee, might you be willing to eventually try making an Espresso brewer despite its potential difficulties?

No, I don’t see that happening. I’d rather have Mugsy do one thing extremely well than do lots of things only passably. Also, the pour over method fits in a way that other types of coffee do not. It’s good enough for coffee snobs to get excited about and it doesn’t scare off the people who just want a no BS cup of coffee.

The best thing about Mugsy is that it is fully open-source, and therefore can be driven by its community. How do you envision the community to interact with each other online, whether it be over the internet, or through Mugsy?

Initially we will be doing the standard Github open source project workflow for the main Mugsy runtime, with an improved developer forum and in-depth docs. But that will expand to other parts of the system. I see a huge opportunity for aftermarket models/parts to change the way Mugsy works at a structural and mechanical level. As an example, there’s a coffee maker called an Aeropress that is basically a manual vacuum press for espresso. I designed a conversion mount that when combined with a linear servo or two, you can use an Aeropress within Mugsy and you get all of those same benefits. So as previously noted, a stock Mugsy will always stick to the one function done perfectly philosophy but the user is free to hack away and share those improvements or additions with everyone. And since it’s running a full Linux stack the integrations and software possibilities are completely open ended. Whether that means a Spotify integration or a full on IOT home controller that brews when you get out of the shower, it’s really up to what people imagine and dream up. There will definitely be a simple interface for users to add community apps or plugins although that is not yet fully fleshed out.

Benchmarking pump voltages for the @heymugsy water flow rate data set. #drinkcoffeeanddestroy

A post shared by M. Argyle (@marrrgyle) on

In your reddit post, you say that you like to “pull the curtain back” on expensive consumer goods, in which you effectively obliterated the $1500 price tag on higher-end coffee appliances. What are some other notable examples of this that have interested you?

What got me started on that path was designing a smart outlet. With a relay, an ESP8266 or similar and a hacked-up phone charger you can beat the pants off most of the consumer options for a few bucks. I made that a couple of years ago and prices have come down a bunch so it’s not a huge disparity like Mugsy, but it was the first project that got me thinking about it as a design philosophy. I am currently working on some things in the audio recording space with a similar goal. Markets that have a large “Prosumer” user base are perfect for this type of work as the better gear usually uses similar spec parts as the entry level stuff but with more thoughtful construction or usability.

How many hours of time do you think you’ve put into Mugsy so far? With that said, do you prefer the ‘voyage’ or the ‘destination’?

I couldn’t say how many hours but I’ve been working on it seriously for about 3 years. It’s been a long slog for sure. For me the voyage is really fun because I tend to only work on things I don’t yet understand. There’s this built in reward system where I’m having these small wins really frequently as the project progresses. But that’s something only I can see; to everyone else they just see the end product and it not being done. So it’s really tough but you have to not be fooled by the joy of the process and you need to get it done and ship. And it gets harder the closer you get to the end, the hurdles are always scaling up as you’re moving forward.

A post shared by M. Argyle (@marrrgyle) on

If the first 100 test kits go well, how likely can we expect another manufacturing run?

The initial run will be for 100 units, kit details are still coming together. Up until World Maker Faire and some earlier feedback I wasn’t even planning on kits.  But it was easily the most frequently asked question by a large margin and it’s a no brainer now, so I will be releasing kits at the same time but I don’t have quantity or cost details just yet. With that said, if this initial release goes well, I have plans to scale up production and continue moving Mugsy forward.

On a scale of 1 to Professional DIYer, how difficult will it be to put together one of these kits?

I think anyone who has some basic computer knowledge, a desire to tinker with things and experience building some Ikea furniture will have no problems building the kit. You won’t need to know how to solder or write any software to get Mugsy up and brewing. People who have more experience will be able to get in there and really make it their own but that is not a requirement to get to the baseline.

Is there anything else you’d like to say about Mugsy and share with your following?

Only that if you have an idea, it’s never been easier to make it happen. You don’t need any esoteric knowledge to get started and figure it out on the way. There is a lot of gate keeping because people work really hard and have fragile egos, but the reality is anyone can do it if they have the desire and most importantly the follow through. Advancing tech is not just for the folks already on the top of that mountain, it’s open to all of us.

A post shared by M. Argyle (@marrrgyle) on

You can get updates to Mugsy by visiting and subscribing, or by following Matt Oswald on his Instagram @marrrgyle or Twitter @mattargyle. Mugsy will sell for around $250-300 for the first 100 test kits, according to Matt himself. I hope you guys enjoyed reading up on his responses to Mugsy, and I’m very much looking forward to getting one myself!

Share on Facebook
Tweet about this on Twitter
Email this to someone

Feel free to share!