A Soylent Mixing Machine & Business Model


As a high-end kickstarter tier, I’d really love to be able to buy a soylent mixing machine. Essentially, a robot to further remove me from the preparation of soylent, but allowing me to observe and modify it to conform to my shifting dietary needs.

I’d like an app that allows me to create, edit, and share variations on @rob’s original formula using a web browser or mobile device, so I can find my ideal formula if necessary.

When I run low on ingredients (and I should run low on them all at once), I’d like a replacement cartridge ready and waiting to be sent to me via the soylent website. I’d like to be able to subscribe to this service and constantly receive new cartridges for my soylent robot as needed.

Then, I’d like to be able to 3D print my own soylent mixing robot so I can put them in areas with high concentrations of homeless people.

Stretch goal? :smiley:


As a ex-mechanical/electrical engineer, I would be more than willing (and would love too) to do all the CAD/design work for the Soylent Robot Maker. I do have plenty of experience with micro-controllers to control this machine.

As far as 3D printing, doubt that will happen since we can’t even print complicated mechanical contraptions cost-effectively.

So, you click “high protein morning drink” on your tablet/phone app and your robot pulls down the recipe for that mixture from either a local or web database and in a few seconds, the robot is away making your morning meal. Getting pretty out there, but surely doable…


Cool! I’d love to see your designs. In my mind, it’s made of individually sized, gravity-fed containers that all pour into a common mixing area.

It seems like much of this is just plastic containers, which could be 3D printed. You’d still need to buy the Soylent robot brain in order to mix anything, but it could be an effective way to cut costs or reduce Soylent’s participation on the overal supply chain (so it can focus on the mixture itself).

Anyhow, very excited about the potential of Soylent, and looking forward to further discussions.


Well, as there are around 40 different elements, you would have to have very specific containers with very very sensitive strain gauges (weight sensors). Your “plastic containers” could not be 3D printed, otherwise, you wouldn’t know the exact weight, and your measurements would be off.

You might, MIGHT be able to get away with a couple of measuring devices that you poured into, but your pour control would have to be crazy fast and accurate.

But, if you had all that, you could easily give the measurements in whatever format you wanted (grams, ounzes, etc) and have them calculated out by the micro-controller to be measured out. I would think an arm chip would be required for this sort of processing power. Something much faster than an arduino/netduino. Might as well build in wifi and a storage/tracking system to track your daily usage and intake and other stuff.


You seem much better equipped to make informed choices about this than I am, so I’ll defer to your expertise. Let me know if you come up with a design… I’ll buy one! :smile:


As an electrical engineer, I’ve thought a lot about this. I would especially like something that would compensate for the regular food you chose to eat. For example, if you went out for steak on Sunday night, monday’s formula would cut back on protein and fat. Then as long as your ‘leisure meals’ were within reason, and you drank soylent twice a day, you would still be guaranteed a perfectly balanced diet.

The trace vitamins and minerals are trivial to measure and supply ahead of time as a standard cartridge, and the body can store and purge excess easily. The water-soluble nutrients and macronutrients are the real problem. However, there are few enough that a counter-top device could measure them precisely. Look at auger fillers for the industry standard mechanism.


@rob I’m really excited to hear you’ve been thinking about this, and that you’re an electrical engineer. I’d really love to see the Soybot (or whatever you decide to call it) in my kitchen, so let me know if I can help in any way. I don’t have a whole lot of relevant experience, but I have a pretty diverse array of contacts, and I’d be happy to mine them on your behalf.

I think what you’re doing is capital-G Good, and I want you to succeed.

Updated: You know what. Actually, I could totally produce, design and build the web app component of this for you. I have 18 years of web and game development experience, and I’d love the challenge. Contact me at craigotronprime@gmail.com if you’re interested.


I had an idea that you could have 3 or 4 or more sized packets of product and you could mix and match packets for a near perfect match. Mixing your own is great but not a good business model. Unless you had a big tub like body building products.


Hi guys,
Just passing along some developments in the world of ‘mix-these-things-together’ home brew tech,
the BarBot idea was what I thought of when reading this thread, The idea being people develop processes to measure (accurately) and mix multiple different liquids, and produce a final drink.

Yes, Mobile and Tablet applications have been developed.
Maybe there’s something in the field out there that could be adapted to the ideas you’re looking for? If the components could be shipped/stored in liquid form, then mixed via an automated mixing bot.

Tested.com’s youtube channel have a good general overview of the 2013 meetup


Rob, if you ever wanted to get together and work on a design to set up something like this, let me know. I did prototype engineering work for 4 years during college and some after.


RE: Different elements in specific containers and sensitive strain gauges.

I don’t think that’s the case. Most of the micro nutrients and elements are going to be pretty much universal, so a single “dose” will be common for everyone. The things that vary based on sex, body type, and genetic predisposition can be be placed in supplemental containers (e.g. a women’s supplement that adds iron, folic acid, etc. above the normative dosage).

Also, you wouldn’t need minute measurements because that can be done in manufacturing by “cutting” the nutrient with a filler. You then standardize on the filler (e.g. rice flower) so that you have a known nutrient load embedded in the cutting agent and the bulk items (carbs,protein, fiber) can be adjusted automatically by the system based on how much cutting agent is used in the other cartridges.

So, you’d go to the “soylentbot.com” website, sign in and fill out a questionnaire and based on those answers, you would get your SoylentBot with an initial formulation for a week or two of use. If you experienced a problem (e.g. fatigue) then your order would be adjusted (e.g. more iron). You could also submit blood work panels and the system could use these to make adjustments to your orders as @rob did by hand. Further problems (or even benefits) could be submitted either between orders if it’s an emergency, or at time of reorder, thereby collecting data that could be used by a nutritionist and/or dietary physician to make adjustments to the expert system that makes adjustments to your orders, and to provide new products to meet special dietary needs (e.g. epileptics’ and diabetics needs to avoid carbs, possible allergies to certain ingredients, etc.).

Such a system would even be invaluable to for research purposes, allowing volunteers’ dosages of certain items to be adjusted under the supervision of the researchers, and their information collected using the same interface as the general questionnaire.

The problem with all this is I think is you’re going to run afoul of the FDA. This is especially true if you follow through on “prescribing” and “diagnosing” as I propose in the third paragraph. Expert systems have proven themselves to be much better than general practitioners in diagnosing disease within a narrow area of specialty (e.g. recognition of cancer based solely on reported symptoms), but nobody to my knowledge has ever been able to produce such a system commercially to give guidance to patients without running afoul of the FDA. I’d be interested in knowing if such a case exists. The solution for this may be to have a dietary physician involved in design and implementation and submit each generic “cardridge” formula as a basic dietary supplement which by passes the “safe and effective” rules of the FDA. Any “custom” cartridges, however, may require a prescription and each item in them may have to go through clinical trials, making it ridiculously expensive (and probably taking that option off the table).

I’m sure once you’re up and running, you can hire experts on this process to help walk through it. Considering the number of vitamin manufacturers in the US, the hurdles for supplements sound very doable, and the bulk items like whey, etc. should be no problem at all since they are clearly and recognizably food ingredients.


To offset the FDA responsibility, you could move towards a suggestion type, such as, here is the basic list/measurements [listA], then with a list of changes such as, feeling tired? add more iron, recommended amounts are [x] and then the user takes responsibility of changing their SoylentBot through and easier to use interface system. Then users can upload their own mixtures to a community type site (which is already under work right now) and share what they changed and why and the benefits and draw backs.

All of this API would be exposed to the SoylentBot so you can push recipes up and down through the web app and make fine tuned adjustments.

As for the combination of different ingredients, the downfall to that is that you can’t fine tune your recipes because you already have a predefined mixture. The idea of this system is to allow an exact recipe for you, so you can change your own mixture to what suits you best.

Anyway, this has a ways to go, but I hope to keep this discussion open and hopefully get rob more involved once he has time and getting is more knowledgeable experience



If Soylent is considered a “food” and not a diet or supplement aren’t the regulation oversight and approval requirements diminished?


I don’t know about diminished, but they are different.


If the device we’re designed so that the ingredient containers were open and general, then you could route around any possible FDA issues by making the device double as a spice-mixer. Just replace the soylent cartridge with one filled with various spices, and sell the soylent cartridge as a supplement pack or something.

An internet enabled spice mixer would be a pretty useful device on its own. Having it double as a soylent mixing robot would be a great way to introduce “normal” people to soylent.


Not a bad idea… Having a “mixing robot” that can work for, multiple scenarios could easily bypass a lot of regulations.

I guess I should start drawing up some designs.


Would love to see the Soybot come to fruition and help out in any way I can. Although I’m a EE major in college I’ve done a lot of work with device drivers and software. @prefinem, perhaps we should have a hangout for a team of engineers interesting in pushing forward this idea on a weekly basis?


Sounds good. Here is a group for just the Soybot: https://plus.google.com/communities/118345161414277136767

Once we get serious (and have some valid ideas), I will open up a project on openmanager.net and we car start designing it there.


It’s really cool to see enthusiasm start to form around this idea. I love the thought of a world where a little spinning robot in our kitchens and corner stores keeps us all healthy.

@prefinem do you have any experience with the IDE/Game Engine, Unity? As you come up with CAD examples, I’d like to try and port your designs to Unity to create a working software object that we can distribute widely for “pre-alpha” testing. User feedback can drive iteration of the design.

Would you consider starting a github repository for your work, that we can all add to freely?

@rob Objections, thoughts, and guidance are greatly appreciated, always.


@craig, I haven’t really worked with Unity (or any game engines) before. But honestly, I use SolidWorks and will be able to do simulations on it for flows, stress, movement, etc… although I would be happy to port anything over for you. I can upload to my GitHub account when I get some things going. I would ask you join the google group if you want to have more input, etc…