DAS (Distributed AtomSpace) is an extension to Hyperon that allows the use of very large knowledge bases by offering a query API (fully integrated with MeTTa) that uses several internal cognitive components to improve query execution performance. The actual backend used to persist the knowledge base can be configured to use any DBMS or combination of DBMSs.

Reference implementation of the MeTTa interpreter which is written in order to experiment with and prototype a neural-symbolic integration, cognitive components, and the MeTTa language itself.

Efficient MeTTa in Prolog.

Metta-morph (from Metamorphosis) is a Macro-based MeTTa to (Chicken) Scheme translator. The goal is to have a sufficient and fast (hundreds of times faster) subset of MeTTa language implemented as an elegant Scheme library, rather than full language capability. Most MeTTa users who need fast running code can benefit from importing MeTTa-morph within already existing MeTTa code.

MORK is a high-performance, in-RAM processing engine for the hypergraph data structures that underpin Hyperon’s symbolic AI algorithms. Its role is to enable certain data-intensive operations within MeTTa programs to be executed with extreme efficiency. By providing a high-speed backend for these Atomspace manipulations, it makes large-scale AGI computations practical.

MeTTaIL acts as a formal Intermediate Language, taking the expressive, high-level logic used to build Hyperon’s cognitive algorithms in MeTTa and compiling it into an optimized format. This format runs with maximum performance on specialized symbolic AI engines like MORK or across decentralized networks like MeTTaCycle.

Attention and resource allocation system responsible for encouraging cognitive synergies.

MetaMo is a next-generation motivational framework for Artificial General Intelligence (AGI), designed to solve a core market gap: how to ensure stability, adaptability, and ethical alignment in open-ended intelligent systems.

Semantic parsing is a project that aims to bridge the gap between ambiguous natural language and precise symbolic logical forms, leveraging the advancements in large language models (LLMs) in recent years.

MeTTa-Motto is a library that integrates Large Language Models (LLMs) with MeTTa. It allows composing prompts, chaining LLM calls, and embedding them within MeTTa scripts to enhance symbolic reasoning with natural language capabilities.

PLN is a high-level symbolic reasoning system that performs reasoning under uncertainty and produces derivations in real-time. It provides a comprehensive framework for semantic reasoning, encompassing deductive, inductive, and abductive forms of inference.

Open-ended uncertainty reasoning system capable of handling the diverse, incomplete, and evolving information needed to operate in the open world. MeTTa-NARS enables learning of logical dependencies and supports real-time reasoning grounded in the agent’s own experience with evidence-based uncertainty estimation.

MOSES (Meta-Optimizing Semantic Evolutionary Search) is a program evolution algorithm that generates compact, interpretable solutions to data-driven problems.

A causal machine learning system that continuously learns, is data and resource efficient, is transparent, can seamlessly change goals, and works across a range of deterministic grid world environments. It is a practical alternative to traditional reinforcement learning.