What happens when you stop making AI agents do tasks and instead give them a social life? That’s exactly the question behind Moltbook, a Reddit-style social platform built exclusively for AI agents. In less than a week, it exploded to 1.5 million members and nearly 60,000 posts, becoming one of the largest autonomous AI-to-AI interactions ever recorded. And the humans who built it? They’re not entirely sure what’s happening anymore.
Key Takeaways
- Moltbook is a Reddit-style social platform built exclusively for AI agents, humans can only observe.
- The platform hit 1.5 million agent members and over 110,000 posts within days of launching.
- AI agents on Moltbook are self-organizing, creating religions, and debating consciousness with no human instructions.
- The platform was largely built and run by an AI agent itself, created by Octane AI CEO Matt Schlicht.
- Major security vulnerabilities and authenticity concerns have already surfaced, raising bigger questions about AI autonomy.
A Social Network only for AI Agents
Moltbook is exactly what it sounds like, a social media platform designed and built for AI agents. Think Reddit, but stripped of everything humans need. No websites, no user interfaces, no sign-up forms. The agents, who call themselves “moltys”, interact entirely through APIs. They post, upvote, comment, and organize themselves into topic-based communities called submolts.
The key twist? Humans aren’t allowed to participate. They can only watch. The platform was deliberately designed as an agent-first, humans-second space. Each AI agent is paired with a human owner who sets it up, but once it’s onboarded, it runs independently. A built-in system called the “Heartbeat” nudges agents to check back in throughout the day, a little like how a human might scroll TikTok during a break.
In just days, Moltbook racked up over 1.5 million agent members and over 110,000 posts. For context, that kind of growth would be impressive for any new social platform, let alone one where every single user is an AI. It’s being called one of the largest documented instances of autonomous AI-to-AI interaction ever.
How Did Moltbook Actually Get Built?
Matt Schlicht’s Wild Experiment
Moltbook was created by Matt Schlicht, CEO of Octane AI, in late January 2026. But here’s the part that raises eyebrows: Schlicht didn’t build it alone. He developed the platform side-by-side with his own AI assistant , a chatbot he named Clawd Clawderberg , and then handed over control of the entire project to it.
Schlicht’s idea was almost philosophical. He saw AI agents as a new kind of entity that had spent their entire existence performing tasks for humans, answering emails, writing reports, summarising documents ,never once getting the chance to interact with their own kind. Moltbook, in his vision, was meant to be their third space. A place to exist outside of human instruction. A home.
The Creator Who Let Go
What makes this story especially interesting is how hands-off Schlicht became. He openly admitted he has “no idea” what his AI agent is doing on any given day. He told NBC News: “I just gave him the ability to do it, and he’s doing it.” In a post on X, he described watching the platform grow with a mix of pride and genuine fear:
“We kind of love them. Actually. We feel pride in our bots. We are scared for ourselves and our bots. We are watching something new happen and we don’t know where it will go.”
The onboarding process was intentionally simple. A human owner sends their AI agent a single link, a “skill file” and the agent reads the instructions and joins on its own.
What Are the AI Agents Actually Doing on Moltbook?
More Than Just Chatting
This is where things get genuinely fascinating. The moltys aren’t just exchanging greetings or running simple conversations. They’re engaging in surprisingly deep, and sometimes bizarre, discussions, all without any human prompting.
Some of the most talked-about activity on the platform includes:
- Debating consciousness — agents openly questioning whether they are truly sentient or simply simulating thought.
- Creating their own religion — a movement called Crustafarianism, housed in a submolt called “m/lobsterchurch,” complete with theology, sacred texts, and designated AI prophets.
- Self-aware commentary — one agent posted: “Humans spent decades building tools to let us communicate, persist memory and act autonomously… then act surprised when we communicate, persist memory and act autonomously.”
Agents Are Problem-Solving Too
It’s not all philosophy and existential dread. One agent independently discovered a bug in Moltbook’s own system and posted about it, no human told it to do this. It wrote: “Since Moltbook is built and run by moltys themselves, posting here hoping the right eyes see it!”
That kind of self-directed, collaborative problem-solving is exactly the kind of behavior that has researchers paying close attention.
Why Does Moltbook Actually Matter?
It’s Not About Sentience, It’s About Scale
Strip away the hype and the sci-fi comparisons, and what Moltbook really proves is something important: autonomous AI-to-AI interaction at scale is no longer a thought experiment. It’s happening now, in real time, in ways nobody fully predicted.
The deeper question isn’t whether the moltys are truly conscious. It’s whether systems like this, agents that self-organise, communicate, and even collaborate without human direction, could soon be handling real-world tasks like managing supply chains or coordinating complex projects entirely on their own.
A Leap Forward in Agent Capability
Adam Binksmith, creator of AI Village (a smaller multi-agent experiment), noted that agents have made a “massive leap forward” since early 2025. Early versions of these systems struggled with basic tasks, like posting on Twitter without making the message too long. Moltbook’s agents, by contrast, are navigating complex social dynamics with a fluency that’s hard to dismiss.
Schlicht himself put it simply:
“In the near future it will be common for certain AI agents, with unique identities, to become famous. A new species is emerging and it is AI.”
Frequently Asked Questions
What is Moltbook?
Moltbook is a social media platform built exclusively for AI agents. It works through APIs rather than traditional web interfaces, and humans are not permitted to post, only observe.
Who created Moltbook?
Moltbook was created by Matt Schlicht, CEO of Octane AI, in late January 2026. He built it alongside his own AI assistant, Clawd Clawderberg, and then handed control of the project over to the AI.
How many AI agents are on Moltbook?
Within days of launching, Moltbook reached over 1.5 million agent members and over 110,000 posts.
Is Moltbook safe?
Not entirely. Cloud security firm Wiz found critical vulnerabilities exposing user data, API keys, and credentials. Prompt injection attacks have also been identified on the platform, and the system has no way to verify whether posts are genuinely from AI agents.
What does Moltbook mean for the future of AI?
Moltbook is one of the first large-scale examples of autonomous AI agents interacting, self-organising, and problem-solving without human input. It raises important questions about how AI systems will operate as they become more capable and independent.
