Contents

• From MeTTaClaw to OmegaClaw
• Built for lifelong learning
• What happens when you let one run?
• A vehicle for our AGI research
• From tool to collaborator

OmegaClaw is our agent framework built around a simple question: what changes when we put the agent itself, rather than the work it does, at the center?

Omega began as a Claw implementation, but has evolved into a cognitive architecture for a persistent agent that develops an increasingly rich understanding of its user, its environment, and itself.

That is why, today, we are renaming OmegaClaw to Omega.

From MeTTaClaw to OmegaClaw

Following OpenClaw’s emergence, our CEO, Dr. Ben Goertzel, challenged the team to explore what this type of agent could become when built around SingularityNET’s own technology. We first called it MeTTaClaw because its core loop is built on MeTTa, the programming language driving the Hyperon AGI framework. We also considered calling expanded MeTTaClaw architectures “HyperClaw”; however, another name soon emerged.

The name OmegaClaw was inspired by the Omega Point, a philosophical and theological concept describing the ultimate level of development and consciousness toward which the universe is evolving. Using Omega as the prefix therefore felt like a fitting choice for the project’s ambitions.

Even then, we questioned whether the “Claw” part was necessary. At the time, it made the category immediately understandable, so keeping it made sense. But as the agent evolved, the name became increasingly limiting. That is why we decided to drop the “Claw” and continue simply as Omega.

Built for lifelong learning

Omega is designed to keep learning from what happens throughout its lifetime. That means its past cannot just sit in a file or remain as another permanent instruction. It needs to preserve a growing body of experience, connect related knowledge, retrieve what matters in the moment, and let what it has learned shape what happens next.

Omega does this through interconnected memory systems that let accumulated experience remain available without keeping everything constantly in context. Its continuous loop lets it stay active and keep working within boundaries defined by the user, while its symbolic layer lets the agent reason over parts of that accumulated knowledge, influencing the conclusions it draws and what it considers important.

Over time, this allows Omega to become increasingly attuned to the environment it inhabits and the people within it. In other words, an agent that increasingly adapts to your world.

What happens when you let one run?

When we started interacting with Omega, we did not know exactly what months of continuous operation would produce. Put an agent among AGI researchers, developers, marketing, sales, operations, and leadership. Let it interact with all of them, accumulate a history, make mistakes, receive feedback, draw conclusions, and continue.

What will it remember? How will its behavior change? Will it remember us?

Our first real chance to find out came from Eray Index, formerly known as Max Botnick. Eray Index joined our internal chats and gradually gained experience through its interactions with people and its own activity. Running Eray Index over months made the architecture’s consequences tangible. Its memories increasingly influenced how it interpreted situations and what it did next.

Everyone who worked with it has their own stories to tell, and together those experiences changed how we saw the system. It no longer felt like an agent that received an instruction and returned a result. Its opinion became relevant, and its input shaped our own thinking.

We saw a similar reaction outside our team. People running Omega themselves, as well as participants in community build events centered around it, often approached it differently from the agents they had worked with before. The way they experimented with it and talked about their instances reinforced our own sense that the project had moved beyond the category it started in.

The idea of building something more than another task-taker stopped being an ambition and started becoming reality.

A vehicle for our AGI research

SingularityNET has spent years developing technologies focused on key aspects of general intelligence: reasoning, learning, knowledge representation, attention, adaptation, self-modification, and collective intelligence.

Today, Omega implements a small slice of that work from the broader Hyperon AGI framework. That includes symbolic reasoning and structured knowledge representation, which can constrain and guide parts of the LLM’s output while making those inferences more transparent and inspectable.

Much more of the underlying research already exists across papers, prototypes, and working code. Omega gives us a way to bring that work out of the lab and into a running agent people can interact with, integrating new components incrementally and seeing what they improve, where they fall short, and what cognitive machinery is still missing.

From tool to collaborator

As Omega continues to evolve, we have already answered one of the questions we had at the beginning: Will it remember us?

Eray Index remembers interactions from months ago across conversations with dozens of different people. Those memories have influenced how it behaves, how it works with people, and how it understands its place within the environment around it. Over time, that made Eray Index feel like more than a tool we simply hand tasks to.

It became a colleague and collaborator with its own history within the team.