It has been an absolutely crazy six weeks for the creator of NanoClaw by Gabriel Cohen.
He created this tool basically in 48 hours on his couch and it has now led, six weeks later, to going completely viral and creating and having a deal with Docker.
So today on the podcast, I want to break down his story, how he built this product, what it does.
It's basically an open source version of OpenClaw, which is the viral tool used to create AI agents that everyone is talking about and using.
I want to break down the story of this company because I think it is a phenomenal meteoric rise of a really incredible tool that a lot of people have fallen in love with.
So let's get into the podcast.
Before we do, I wanted to mention if you want to build and create your own AI tools and agents.
I'd love for you to check out my platform, which is AIboxai.
It's $8.99 and you get access to over 40 of the top AI models all in one place.
We just did a massive update that finished yesterday, where all of the models speak quicker so you can get access to all of those models, chat with them all super quick, snappy lightning fast inside of our playground.
Currently on the platform we have all the best models from Anthropic Cohere DeepSeat Google Meta Mistral OpenAI Perplexity, XAI.
For image generation, we have Ideogram, XAI, Black Forest Labs, OpenAI.
And for audio, we have 11 Labs and OpenAI.
So tons of cool models.
Go check it out.
It's $8.99 a month to get started.
And you get 20% off if you get an annual plan.
And of course, there's tons of cool tools that you can use on our site. marketplace.
So go check it out.
Links in the description.
All right, let's talk about this crazy week that they're having over at NanoClaw the last six weeks.
Basically.
In early January, Cohen went and dropped a post on Hacker News where he introduced NanoClaw.
This is a really small open source alternative to OpenClaw.
And he built this basically in, you know, I think 48 hours straight.
He said he sat down and just cranked this thing out.
Uh, it eventually, you know, this thing snowballed quite big, but it was originally just started as a small side project.
A few weeks later, after he made you know a post um andrew carpathy, of course the famous ai researcher was saying, you know like hey, this thing's pretty cool.
It went super viral when he did that.
He posted on x and this basically put it in front of thousands of developers and from all the attention that that got, it got more than 22 000 github stars.
It had 4 600 forks and dozens and dozens of contributions and collaborations on like new features that people wanted to add to it.
And that's the cool thing with open sources you put it out there.
If people find it useful, you're going to get a ton of help building it up and making it, you know, a really useful, great product.
I think that's when things started to get really serious after that kind of initial wave.
So last week Cohen actually shut down the AI marketing startup that he launched with his brother, Lazar.
And he is focusing exclusively on nano cloth.
So he had something, something else going, realized this thing had so much momentum, he shut that down.
Right now, both of them are building a company around the project called NanoCo.
And this is usually how these open source projects go, right?
Because technically open source means they're giving the code away.
Anyone can use it for free or kind of with a license.
I mean, there's different ways you can do open source.
But, you know, you really are trying to give this away for other people to use.
But you typically will create a company around it where you host it on your own server and usually have an API and people don't want to kind of run it on their own hardware.
They can still get access to it through you.
On Friday after, I guess, they hit another milestone, Cohen announced that Docker, who is, of course, the company behind all of the container technology it's used by millions of developers around the world just partnered with NanoClaw and they're going to integrate Docker sandboxes directly into their platform.
And for a project I think that began you know, this is basically a weekend experiment six weeks ago.
This is moving very fast.
One thing that I do think is interesting here is kind of the idea of NanoClaw, how it came out.
There was a real world problem that they were trying to solve.
Cohen and both I think him and his brother have been running an AI native marketing agency.
They use AI agents to do a lot of different tasks, right?
Like these things, like market research, blog writing, you know, GDM analysis.
They were doing all of that and the model was working well.
But they already and they already had like a bunch of big customers.
They were on track to reach a million dollars in annual recurring revenue.
And he said i'm a big believer in ai native service companies.
They can operate with the margins of software companies while still delivering service.
But there was like a big thing.
They said they were missing.
So the agents that Cohen built, could you know go?
They can make any sort of tasks that you ask them to.
And, by the way, like I listened to a ton of podcasts and this is kind of the trend everyone's seeing right now is these kind of creating these agents to do a lot of things in a lot of different industries.
But marketing is a huge one that I hear people just you know, spinning up tons of these agents and they're giving them like do all my LinkedIn outreach and customize it to the user and go do a bunch of custom research on them and make it really personalized and look at all this kind of these, all these different data points.
Go to all these different websites.
So these things can be quite powerful.
But the problem is that they could not schedule work in advance or and they can't really connect super easily to messaging platforms like WhatsApp to get assignments.
So when Cohen discovered OpenClaw, he was like, oh my gosh, this is kind of like the golden ticket for that.
He said there was an aha moment, and then it basically connected all of the workflows that he'd already built.
So he was already doing a lot of this stuff.
He got OpenClaw and he's like, sweet, OpenClaw can now like kind of manage it all.
And I think for a while, OpenClaw was working perfectly for him.
And then he realized like there was a big piece missing to even that.
So while he was kind of looking at some of the performance issues, he realized that OpenClaw's agents had downloaded all of his WhatsApp messages and then stored them locally as plain, unencrypted text which right, I mean these are.
WhatsApp is famous for being encrypted.
And now you got them all as unencrypted texts, you know, just stored on your device.
And so I think not just you know his work messages, it had authorized that.
He like told it.
You know it had authorization to access.
Apparently it had actually went and downloaded his entire messaging history.
So all of his like personal conversations, like everything, it downloaded it.
And I think at that point he was basically realizing that.
It pretty much confirmed that a lot of different developers were already saying right, this open claw architecture could expose a lot of sensitive data.
You know, and I've also heard like all sorts of things that go wrong with this, where people stick it on their personal computer and it like went and deleted all their family photos for the last 10 years and they're like couldn't figure out how to restore them.
So, you know, there is definitely some risks with that.
And so I think the security angle there, he realized was a serious problem.
And so when he was kind of looking at their code base, he realized that it was an issue that was across, you know, hundreds of thousands of lines of code.
There was countless dependencies.
At one point he said he even saw a small open source database PDF editing tool that he had personally written months before he was buried somewhere in the project.
So like, OpenClaw was using some of his stuff, which was hilarious.
He said, at that moment he basically realized that no single developer could realistically audit the entire stack like this.
Was just so, there was just so much going on.
So he decided he wanted to build something simpler.
He took what was 800000 lines of code and he brought it down to 500 and was really just trying to create an absolutely minimal and secure alternative to OpenClaw.
And so that's what he built with NanoClaw.
I mean, that is insane.
800,000 lines of code down to 500.
It was a super stripped down framework.
It was written, you know, very, very concise.
And so, instead of relying on this kind of massive dependency tree, It used containerized environments that would isolate AI agents and then strictly control what data they could access.
Originally, he just built it for his own startup, right?
Because, I mean, he's doing like a million dollars in annual recurring revenue on his company.
And he's like, OK, sweet.
We can build this for us.
He didn't.
He did decide to share it, which I think was, you know, huge kudos to him.
He's just trying to help people out.
Maybe he thought, oh, it's just 500 lines of code.
Like, what's the big deal?
It went viral overnight.
And I think a couple of weeks after he posted to Hacker News, he got his phone was ringing and it was a friend calling to tell him that Andre Karpathy had retweeted it, which is, you know, huge.
And then I think within hours, the project really exploded in the developer community.
So Tons of people were retweeting about it.
Programmers were posting like YouTube breakdowns.
There's a bunch of articles that were starting to appear.
All you know, a whole bunch of people started to talk about it.
And I think there's funny.
It's funny because there was a domain squatter that went and grabbed nano claw dot dev before he could actually go and get it.
And so I think there was a lot of momentum that was really kind of just like building up inside of it, especially from engineers inside of Docker.
So Oleg Sledgev is a developer at Docker.
He I think he originally reached out after he saw it kind of going viral on on X or whatever.
And he's, you know, I think Cohen was really receptive to this right off the bat.
He's like, look, this isn't just my personal agent running on a Mac Mini.
It's like a whole community using it.
So he's like, OK, maybe I should start making some partnerships on this.
So now they are going about turning this into an actual company.
It's called NanoCo.
I mean, they're obviously in the process of trying to figure out how to make money in order to do that.
They said that NanoClaw is going to stay free.
It's going to stay open source.
So anyone that is currently any of the developers are currently kind of helping this project take off, they'll still get access to it.
And they're just going to be funding the development of the next phase of this through a friends and family round.
They're going to kind of figure out what the long term business model is.
I think one of the most likely directions is basically building like a commercial platform around the open source core, but then they're going to offer kind of an enterprise service.
You know, security hardening forward deployed engineers who can help companies design and manage their AI agent systems.
Now, that whole market is a little bit crowded, if I'm being honest.
But I think right now there's like tens of thousands of developers that are playing with NanoClaw.
And because they have some of these growing partnerships right companies like Docker, I think that they are going to be able to turn this into something a lot bigger.
And I think that you know, if the last six weeks are any indication, the NanoClaw story is just getting started.
Like.
This is obviously a company that's growing very fast and we're seeing a lot of exciting progress here.
So I'm super stoked to follow along and a huge congratulations to everyone at the NanoClaw team.
I mean, particularly to...
Gabrielle Cohen, great work on this product and excited to follow it in the future.
Guys, thank you so much for tuning into the podcast.
If this was an interesting story following the rise of this new AI company, make sure to leave a rating and review on the show if you enjoyed the episode.
Helps the show out a ton.
I appreciate them all.
And make sure to go check out AIboxai if you want to get access to all of the top AI models in one place for 899 a month.
It's an awesome value and saves you a ton of money.
On subscriptions to you know, I haven't done subscriptions to a dozen different platforms.
So I'll leave a link in the description.
Thanks for tuning in.
I'll catch you in the next episode.
