Hyperon Index

Draft — This content has not been approved for publication.

Welcome to the Hyperon Index, a curated technical document designed to provide an intuitive and demystified understanding of our AGI frameworks and their constituent parts. Hyperon is SingularityNET’s Artificial General Intelligence(AGI) technology stack building on decades of AGI research. Hyperon provides a unified platform for integrating diverse machine cognitive processes — from symbolic reasoning and probabilistic inference to neural learning and evolutionary search.

Much of the significance of the present Hyperon effort lies in the deliberate rebuilding of infrastructure so that these modes of cognition can interact at far greater scale, concurrency, and semantic fidelity than prior generations allowed.

This document serves as the primary reference for our internal R&D initiatives, offering high-level, current descriptions of each component alongside links to demos, peer-reviewed publications, repositories, and technical documentation for those seeking deeper immersion. The result is not merely a taxonomy of components, but the emergence of a common cognitive medium in which learning, reasoning, attention, motivation, and program synthesis can enter into recurrent, auditable loops.

The index is organized into the following key sections (the TLDR):

MeTTa Programming Language: MeTTa is the native "language of thought" — a fundamentally AGI-specific programming language. This section covers its primary implementations, specifically PeTTa, a high-performance interpreter, and Hyperon Experimental, the original reference implementation that established the framework's core principles.
ASI Chain Runtime Environment: The ASI Chain functions as the "blockchain of thought," providing a decentralized substrate for secure computation and cognitive state updates. Critically, this environment is not limited to public networks; it can be deployed on a single machine or a private network of machines for localized usage, ensuring high-integrity, auditable records of all cognitive transformations and transactions.
Knowledge Representations: This section details Atomspace technologies, the representational core of Hyperon’s neural-symbolic approach. In this context, “Atoms” represent symbolic data and formal categories that allow the system to store not just raw data, but the relationships and logic behind it. Systems such as DAS and MORK enable a dynamic knowledge metagraph where code and data are interchangeable, and where neural or sub-symbolic processes can deposit, reshape, and retrieve structured representations through ongoing system dynamics. This allows AI to perform complex queries and self-modifications across a distributed network, treating its own internal logic as a queryable and improvable data structure rather than a static symbolic store.This section details Atomspace technologies, the symbolic foundation of the Hyperon neural-symbolic approach. In this context, "Atoms" represent symbolic data and formal categories that allow the system to store not just raw data, but the relationships and logic behind it. Systems such as DAS and MORK enable a dynamic knowledge metagraph where code and data are interchangeable. This allows AI to perform complex queries and self-modifications across a distributed network, treating its own internal logic as a queryable and improvable data structure.
Hyperon AI Algorithms: Here we describe our core cognitive algorithms authored in MeTTa and executed on the ASI Chain substrate. These algorithms represent the functional "modules" of intelligence: PLN for reasoning under uncertainty, ECAN for managing limited computational resources (attention), and MOSES for creative problem-solving and evolutionary methods. By integrating these specialized processes into a single framework, Hyperon can move beyond simple pattern recognition toward autonomous, multi-stage cognitive synergy.
Cognitive Architecture & Research: This final section covers the higher-level architectural patterns and research directions guiding Hyperon’s development toward AGI, including PRIMUS as well as cross-stack approaches to learning, transfer, memory, and system-wide cognitive coordination.This final section provides an overview of the PRIMUS cognitive architecture, a carefully considered configuration of the layers and components outlined above that SingularityNET believes is likely to give rise to artificial general intelligence.

MeTTa Programming Language

MeTTa (Meta-Type Talk) is a programming language designed to be the native “language of thought” for AGI. It was designed to serve as the central cognitive calculus for the Hyperon AGI framework — a universal glue that allows diverse AI components (e.g. neural networks, probabilistic reasoners, evolutionary models, etc.) to communicate, collaborate, and synergistically integrate their capabilities.

Rooted in principles of both neural networks and symbolic reasoning, MeTTa unifies elements of functional programming (drawing inspiration from languages like Haskell, Idris, and Prolog), logic programming, and dependent typing.

Unlike general-purpose languages, MeTTa was designed to operate natively over cognitive structures — atoms (symbolic data representations), types (formal categories), and transformations — which are stored in a dynamic knowledge metagraph known as an Atomspace. Within this framework, code and data are interchangeable. This design enables:

Interoperability: MeTTa acts as a shared medium and translator for diverse AI systems — a lingua franca for them to not just “plug in” but seamlessly interoperate. It’s a substrate for any species of AI subsystems or paradigms to flow together and combine, allowing their unique capabilities to be expressed, executed, and coherently orchestrated across distributed, interoperable networks.
Concurrency: Leverages a higher-order rho-calculus foundation to treat programs as asynchronous processes that intelligently execute in parallel without blocking. Its systems utilize parallelized backtracking to scale these computations across multi-core and distributed architectures with near-linear performance.
Security and Auditability: Employs a “by-construction” security model to ensure access rights are unforgeable and mathematically verifiable. Within decentralized networks, all state updates are fully transactional and atomic, maintaining a high-integrity, auditable record of all cognitive transformations.
Reflective Self-Modification: Programs can inspect, analyze, and rewrite themselves at runtime. This “reflection” is critical for an AGI to learn, adapt, and evolve its own cognitive processes.
Flexible Reasoning: The language’s structure allows for dynamic type introspection and the programmatic manipulation of its own knowledge and logic.
Nondeterminism/Determinism: Operates inherently as a non-deterministic inference engine, enabling massive-scale parallel search and “lazy” incremental answer discovery across the metagraph. Efficiency is achieved through smart compilers that resolve symbolic data versus executable functions, while low-level kernels allow for explicit, deterministic control-flow in compute-intensive tasks.

Visit Section →

Knowledge Representations

This section details Atomspace technologies, the representational core of Hyperon’s neural-symbolic approach. In this context, “Atoms” represent symbolic data and formal categories that allow the system to store not just raw data, but the relationships and logic behind it. Systems such as DAS and MORK enable a dynamic knowledge metagraph where code and data are interchangeable, and where neural or sub-symbolic processes can deposit, reshape, and retrieve structured representations through ongoing system dynamics. This allows AI to perform complex queries and self-modifications across a distributed network, treating its own internal logic as a queryable and improvable data structure rather than a static symbolic store.

Visit Section →

Hyperon AI Algorithms

Within this section, we’ll review the mechanisms that compose the dynamics of thought itself. Each Hyperon algorithm functions as a specialized cognitive process that animates the system, elevating static knowledge into active intelligence. Expressed in MeTTa and executed across the distributed substrate, each algorithm addresses a fundamental requirement of general intelligence, for example handling reasoning under uncertainty, managing attention and economic resource allocation, and driving evolutionary learning and program synthesis.

Crucially, these are not isolated programs but interoperable modules of a unified cognitive cycle. By enabling these distinct modes of cognition to interact concurrently on shared memory (Atomspace), Hyperon enables a form of cognitive synergy. This interplay of diverse cognitive skills empowers AGI systems to tackle complex, multi-stage problems that would be improbable for any individual mechanism to overcome. What matters most is not the isolated strength of any one algorithm, but the recurrent traffic among them: the dynamics by which perceptual embeddings, attentional signals, rewrite processes, symbolic references, and learned structures continually transform one another through shared states.

The neural-symbolic bridge in Hyperon lies not only in the inclusion of different component types, but in the dynamics by which perceptual embeddings, attentional signals, rewrite processes, symbolic inferences, and learned structures continually transform one another through shared state.

Visit Section →

Cognitive Architecture & Research

This final section covers the higher-level architectural patterns and research directions guiding Hyperon’s development toward AGI, including PRIMUS as well as cross-stack approaches to learning, transfer, memory, and system-wide cognitive coordination.

Visit Section →

ASI:Chain Runtime Environment

ASI Chain is the dedicated blockchain runtime environment for decentralized AGI, serving as the Layer 1 execution fabric where the Hyperon cognitive stack operates. While traditional blockchains like Ethereum function as sequential global settlement engines, ASI Chain is an AI-native worldwide supercomputer architected to handle the massive, concurrent, and graph-based workloads of AGI. Under the hood, this performance is driven by two foundational engines: F1R3FLY, which renders flawless process calculi to ensure exponential scalability, and MeTTaCycle, which compiles and orchestrates AGI workloads. This dual-engine architecture utilizes BlockDAG data structures to allow thousands of non-conflicting AI processes to execute in parallel, breaking the single-file bottleneck of legacy networks.

Functionally, ASI Chain serves as a distributed cognitive substrate — a living medium that connects disparate servers into a single, cohesive network of mind. Historically, it is the first blockchain capable of native inference settlement, meaning it verifies cognitive state transitions (reasoning steps) rather than just validating token transfers. Whether running on a private cluster or the public open network, it provides the secure, immutable fabric where agents, tools, and microservices interact, ensuring that the calculi of consciousness can be composed and executed with the same cryptographic fidelity as a bank settlement.

Visit Section →