Home
Getting Started
The open-source infrastructure for Artificial General Intelligence.
Hyperon is a unified neurosymbolic framework where diverse cognitive processes — symbolic reasoning, probabilistic inference, neural learning, evolutionary search, and attention allocation — interoperate over a shared memory to produce emergent general intelligence.
This wiki is the canonical, continuously-updated guide to that stack. It connects the big-picture ideas to the papers and the working code behind them — a clear on-ramp for newcomers and a current-state reference for experts. Start with a tutorial, go as deep as you like, and never lose the thread from concept to implementation.
What is Hyperon?
Unlike narrow AI optimized for a single task, Hyperon is a composable infrastructure in which many learning and reasoning algorithms collaborate through a principle called cognitive synergy — so the system can tackle problems none could solve alone. Read the full vision & architecture → About Hyperon
How this wiki is organized
- Tutorials — start here. Plain-language explanations and worked examples, with no prior Hyperon knowledge assumed.
- Deep Dives — expert summaries of the current state of each component, with papers, repositories, and implementation detail.
Every topic links to both, side by side — students begin with the tutorial; experts go straight to the deep dive. Neither is gated behind the other.
Your first stops
- MeTTa Programming Language — Hyperon's homoiconic “language of thought,” operating directly over the knowledge graph.
- Knowledge Representations — the AtomSpace metagraph and its high-performance backends (MORK, DAS).
- Hyperon AI Algorithms — the cognitive modules: PLN for reasoning under uncertainty, ECAN for attention, MOSES for program synthesis, MetaMo for motivation, and more.
- Cognitive Architecture & Research — PRIMUS, the meta-architecture that orchestrates the modules into one self-modifying cognitive system.
- ASI:Chain Runtime Environment — decentralized execution for cognitive workloads at scale.
Key resources
- Whitepaper: Hyperon for AGI ⇒ ASI (Ben Goertzel, 2025)
- Reference implementation: hyperon-experimental
- Language site: metta-lang.dev
- Project site: hyperon.dev