← Back to Reference
A comprehensive index of GitHub repositories in the Hyperon ecosystem, organized by role and maturity. Covers the 5 primary GitHub organizations and 200+ local checkouts in the reference collection.
This page is the GitHub-facing repository guide for readers who need a fast orientation to the code surface. For the more curated architectural interpretation layer, see Implementation Families. Mirrors, snapshots, and non-GitHub surfaces are intentionally not all listed individually here.
| Repository | Why it matters | Role |
|---|---|---|
| hyperon-experimental | Main reference implementation of the MeTTa language and runtime surface. | Core language runtime |
| MORK | Most distinctive new substrate work: triemap storage, zipper VM, and reduction kernel experiments. | Knowledge substrate |
| das | Cloud-native distributed memory layer and query surface for Hyperon-era systems. | Distributed storage |
| PLN | Current probabilistic reasoning implementation and rule surface. | Reasoning engine |
| asi-chain | On-chain execution and alliance-facing runtime layer. | Blockchain runtime |
| MeTTa-Compiler | Pure Rust MeTTaTron evaluator and the strongest current compilation path into the F1R3FLY stack. | Compiler / runtime |
| metta-moses | Main modern MOSES-in-MeTTa line rather than the older C++ codebase. | Evolutionary learning |
| custom-atomspace-builder | Concrete domain-facing pipeline showing how Hyperon-style knowledge substrates get used in practice. | Application pipeline |
| System | Primary Repository | Role | Status |
|---|---|---|---|
| MeTTa Language | hyperon-experimental | Core logic engine; Python, C, and Rust APIs | Stable |
| Hypergraph Kernel | MORK | High-performance triemap substrate for MeTTa | Active Research |
| Distributed Memory | das | Cloud-native distributed knowledge graph | Active Research |
| Inference Engine | PLN | Probabilistic Logic Networks for uncertain reasoning | Active Research |
| Blockchain Runtime | asi-chain | ASI Alliance on-chain AI execution layer | Active Research |
| Legacy AtomSpace | atomspace | Predecessor hypergraph DB; widely deployed | Maintained |
| Organization | Focus | Repos |
|---|---|---|
| trueagi-io | Hyperon core: MeTTa, MORK, PLN, Pattern Mining | 18 |
| opencog | AtomSpace, cogutil, NLP, legacy infrastructure | 40+ |
| singnet | DAS, AIRIS, AI-DSL, SingularityNET services | 30+ |
| iCog-Labs-Dev | MOSES, ECAN, bioinformatics, search, MORK SDKs | 49+ |
| F1R3FLY-io | MeTTaTron, MeTTa-IL, Rholang runtime | 10+ |
MeTTa language implementations and evaluation engines.
f1r3node/models protobuf types (no FFI). At HEAD efee4be (2026-01-19, Cargo v0.2.0, 190 commits ahead of the v0.2.0 Git tag), main is 4 months stagnant but feature branches (feature/wam-backend, feature/pln-support, feat/bind-tokenizer) remained active through 2026-03. Per MeTTa runtime cluster pilot Source 3, closed 2026-05-13: [METTATRON-LAZY-NONDETERMINISTIC-CARTESIAN] / [TREE-WALK-INTERPRETER] / [PARSER-TREE-SITTER] / [STDLIB-EXTENDED-WITH-SET-OPS] / [RHOLANG-DIRECT-RUST-INTEGRATION] / [MORK-PATHMAP-HYBRID-PER-OPERATION] / [ASYNC-PER-EXPRESSION-PARALLEL] / [Q-3-DIVERGENT-SEMANTICS]. Active on branches; main stagnant.[INTERPRETER-COMPILER-DUAL-RUNTIME] / [COMPILER-OPTIN-INTERPRETER-DEFAULT] / [EVAL-PROLOG-BACKTRACKING-NONDET] / [STDLIB-CORELIB-PROLOG-HOSTED] / [HYPERON-OPTIONAL-VSPACE-BRIDGE] / [WAM-VIA-SWI-PROLOG-HOST]; quirks Q-1 reproduced / Q-2 fixed / Q-3 reproduced. Default branch vnamed HEAD 74dbe9b8 (NOT master, which is stale/divergent). Contributor forks patham9/ngeiswei/zariuq are stale (ngeiswei shares no merge base — history artifact, not a tradition). Active (S5 closed 2026-06-08).Jetta.g4) frontend + ASM (org.objectweb.asm) backend emitting real .class bytecode; compile-load-invoke REPL; custom JVM Space/Matcher with per-call bindings; gradual types; Kotlin-hosted stdlib. [JVM-BYTECODE-COMPILED-RUNTIME] / [JETTA-ASM-JVM-BYTECODE-PARTIAL] / [JETTA-COMPILE-PLUS-EVAL-EMERGING]; quirks Q-1 not-applicable / Q-2 fixed / Q-3 reproduced. Authors Khudobakhshov + Potapov. main v0.5 HEAD e3849120; active eval-as-runtime work on khud/dev (32 ahead, 2026-06-06). Active (S6 closed 2026-06-08).dev at 3343fbe (2026-03-02); main ~10 months stale; frmlbytecode active to 2026-03-30. IL track, not a direct MeTTa runtime. Active (on dev/frmlbytecode; S4 closed 2026-05-15).language! proc-macro framework that generates, per formal calculus, a PraTTaIL (Pratt + recursive-descent) parser, a Rust-enum AST, and an Ascent (Datalog) rewrite engine. rho-calculus (RhoCalc) is the first-class demo language; carries a real pathmap trie dependency (Adam Vandervorst's crate, scoped to RhoCalc literal/trie ops) and a complete machine-checked Rocq/Coq proof suite (12 .v files, 0 Admitted / 124 Qed). Clean-sheet — no Scala code shared; related to MeTTaIL by GSLT vision, not codebase. Default main at a95726a (2026-05-20); active main + live branches. Zar Goertzel's zariuq/mettail-rust fork (268 behind upstream) carries Lean-export + tree-sitter feature branches. IL/formal-semantics track, not a direct MeTTa runtime. Active (SC-4 Cross-Org Cognitive Repo Sweeps S1, closed 2026-06-09)..gslt theory specs (rho/lambda/ambient); parser-only prototype, no evaluator. [FORMAL-SEMANTICS-SPEC-PARSER]@f1r3fly.io/oslf-editor); standalone authoring tooling. [TOOLING-INFRA]be.adamv Scala metaprogramming/staging library (fork-mirror); the compile-time base under the HDC stack. [METAPROGRAMMING-LIBRARY-FORK-MIRROR]pathmap crate is an actual (scoped) dependency in mettail-rust, and MacroLoop is the metaprogramming base here — but MORK itself is NOT a code dependency of any SC-4 repo (it is a pathmap-dependency + Vandervorst identity-lineage, not a MORK code merge).MeTTa() in-process (binds a space, load-ascii loads biomedical Data/*.metta, runs generated queries) and offers four selectable backends — MeTTa, Cypher (the checked-in app default), a MORK server adapter, and a MORK-CLI path (README-recommended for production). LLMs are auxiliary (title/summary only, not the query backend). Active. [APPLIED-HYPERON-MULTIBACKEND-QUERY-SERVICE]main (24 ahead / 258 behind); the live MORK support lives on the feature/mork-support deployment branch (the one the orchestration submodules). A deployment variant, not an independent cognitive line. [STALE-GENERALIZED-FORK-WITH-MORK-DEPLOYMENT-BRANCH]app/). [ATOMSPACE-BUILDER-API-ON-VENDORED-HUGEGRAPH]rejuve-bio/MORK itself is a server-ops fork of the covered MORK and is out of scope.
iCog-Labs-Dev/bio-semantic-parser; multi-stage NLP (fetch/clean/coref → NER/PubTator/negation → Pydantic+Instructor) emitting structured bio-relations for KG ingestion (a legacy v1.0 stage also generates MeTTa S-expressions via fol_to_metta). [ACTIVE-DIVERGENT-FORK-NLP-BIO-EXTRACTION]Truth_Revision (via lib_pln) to merge per-edge evidence STVs, with MORK-staged proposals and a prompt-routed human-approval gate. [APPLIED-MULTI-AGENT-BIOCURATOR-ON-OMEGACLAW]rejuve-bio/biochatter-metta is a CI-only mirror of iCog-Labs-Dev/biochatter-metta (the iCog tradition; substantive code attributed there, SC-6). BioClaw's OmegaClaw base belongs to the Mindplex “Claw” family (mettaclaw / OmegaClaw / openclaw) as a runtime dependency + naming convention, not code lineage (SC-7). The platform apps AI-Assistant (AutoGen + KG, non-Hyperon) and chatbot-ALIS (Qdrant clinical RAG, non-Hyperon) are out of the Hyperon line. lib_pln loading test are implemented, but the HTN planner and PLN reasoner packages are empty stubs and the generated truth-values are mock/TODO. [HYPERON-HTN-PLN-METTA-WORKFLOW-GEN-PARTIAL-SCAFFOLD]Truth_Revision in BioClaw, while its dedicated planning/reasoning experiment remains a scaffold and its other experiments are non-Hyperon. Cross-org dependencies: iCog-Labs-Dev (biochatter-metta, bio-semantic-parser; SC-6) and Mindplex / OmegaClaw (SC-7).opencog/atomspace imports at HEAD). Same applied-consumer family as the rejuve-bio stack above. The active surface is the demo-PeTTa branch (pin 8fbd847, 2026-06-14); main is stale doc-only (its README points to features/tests/*.metta files absent at HEAD). Hyperon deps = the hyperon pip package + the PeTTa submodule (trueagi-io/PeTTa, the Janus/SWI-Prolog-backed MeTTa runtime); MORK + PathMap appear only as dangling, .gitignored gitlink pointers — not app-used (this is not applied MORK use). Pipeline: Microsoft MIND news-recommendation benchmark ingestion → data.metta (CTR-bucket engagement labels) → an independent native-MeTTa frequent-pattern-miner (ETV truth-values; Apriori/support-style, not the opencog/miner C++ lineage) → a PeTTa backward-chainer wired into a Flask/gunicorn mining + simulation API + AtomSpace-visualizer UI. A standalone MeTTa PLN ruleset (Revision/Modus-Ponens/deduction/induction/abduction, cross-referencing opencog/pln) lives in experiments/PLN/ but is not on the active API path; PLN use stays within the cluster-pilot No-Go boundary (bounded local-chainer semantic queries, not global probabilistic inference). End-to-end partial-but-working (no scaffold stubs, unlike galaxy-htn-neurosym above). Active. [APPLIED-HYPERON-PETTA-RECOMMENDER-MINING-STACK]
Neuro-Symbolic branch is a mixed research side-line — mettaMotto is Hyperon (the metta-motto LLM-agent family, cf. rejuve-bio/metta-motto-biochatter above), while DFOL (differentiable ILP, Gao et al. 2022) + RARL + RuDiK are non-Hyperon neuro-symbolic experiments. Other branches (demo-mork, results, demo-ingestion, ASI-main, SYMBOLIC/Evoda) are variant/deploy/doc surfaces.lib_nal.metta/lib_nal7.metta — Patrick Hammer's NARS/NAL line; embedding long-term memory in AtomSpace format via ChromaDB, remember/query). asi-alliance/OmegaClaw-Core ("OmegaClaw"/Oma) is the larger productized neural-symbolic agent framework on the Hyperon stack (continuous PeTTa loop, channels, agent-managed memory, Telegram/Docker, tested); it carries a real agent-local PLN ruleset (lib_pln.metta — deduction/induction/abduction/revision/modus-ponens, borrowed from trueagi-io/hyperon-pln + opencog/pln) that stays within the PLN No-Go boundary (an applied rule library, not a global PLN system). Lineage: OmegaClaw-Core is a common-root git fork of mettaclaw (shared afbfe9c “Initial commit”, identical early history) that diverged at merge-base 1a7c01e (2026-04-10) into heavy productization (614 commits ahead; current shared-name files have diverged blobs) with active bidirectional cross-pollination (patham9 commits “Updates from OmegaClaw”). [HYPERON-PETTA-CLAW-AGENT-RUNTIME-FAMILY]
BedrockProvider) — no new cognition. [BEDROCK-PROVIDER-DELTA-NON-COGNITIVE] Personal variants leungmanhin/mettaclaw + zariuq/mettaclaw are stale Hammer-line forks; leungmanhin/openclaw + zariuq/vericlaw are a different JS/TS agent tradition (upstreams openclaw/openclaw + godelclaw/vericlaw), not the MeTTa-native Claw line. rejuve-bio/BioClaw (SC-5) builds on OmegaClaw as a runtime overlay. OmegaClaw-Core's home org is asi-alliance (SC-1) — extracted here, cross-referenced (not re-extracted) at SC-1.807bdc17; MIT, Python) — a mature self-improving LLM agent: closed learning loop (skills-from-experience), pluggable cross-session memory, bidirectional MCP client+server boundary, prompt-cache “narrow waist” discipline, single multi-gateway process, subagents + offline batch trajectories, six deploy backends. Zero Hyperon/MeTTa/AtomSpace/PLN coupling (scoped grep at pin; only false positive a “Metta” meditation term) — a workflow to absorb, NOT a Hyperon component; name-coincidence with the Magi HERMES (card 741). nunet/hermes-agent-lite (@7c2c72a1) — NuNet's thin-image + MCP-delegation adaptation: 7 ahead / 298 behind upstream (merge-base 5aec00f7); entire delta = Dockerfile.lite + README + CI (strips Playwright/ffmpeg/TTS/image-gen ~1.5GB→~400MB, delegates to MCP servers). Full workflow write-up: External Agent Workflows+HERMES Agent Workflow. Boundary-excluded: CUDOS (CudoVentures, 60 repos = Cosmos blockchain/cloud-compute) + NuNet distributed-compute/tokenomics/cardano/frontend infra (non-cognitive plumbing); nunet/misc-experiments/f1r3fly-node → SC-4 F1R3FLY. [NON-HYPERON-SELF-IMPROVING-AGENT-WORKFLOW-TO-ABSORB] / [ZERO-HYPERON-COUPLING] / [THIN-IMAGE-MCP-DELEGATION-ADAPTATION].6d395d7b; [HYPERON-RUNTIME-CONSUMER]) is a real Hyperon runtime consumer — hyperon==0.1.6, from hyperon import MeTTa, wraps the LLM-generated query in !(match &self …) + imports the selected BioAtomSpace .metta files + runs it via MeTTa().run() in-process; a from-scratch MeTTa reimplementation of the BioChatter schema-guided-query pattern (emits MeTTa S-expressions, NOT a git fork of biocypher/biochatter). biochatter-metta-server [DJANGO-CHAT-API-WRAPPER] + biochatter-metta-client [SUPPORTING-UI] — the app trio is 2024-era dormant. bio-semantic-parser (@a717ca1d; [ACTIVE-GSE-PUBMED-FOL-TO-METTA-EMITTER]) is the live 2025-11 line: GSE/PubMed abstract + GSM metadata → MedCAT concepts → LLM FOL-like triples → MeTTa list syntax with local regex validation — emit-only, no Hyperon runtime call at HEAD (OpenAI+Gemini providers). Plus bio-data-semantic-parsing [NOTEBOOK-PROTOTYPE] (2025-03 Jupyter precursor) + pubchem2metta [PUBCHEM-RDF-TO-METTA-ETL] (BioCypher MeTTaWriter, no runtime). iCog is the canonical origin; SC-5 rejuve-bio is downstream — rejuve bio-semantic-parser 37/0 ahead (substantive Pydantic-schema/TAXONOMY/retry work), rejuve biochatter-metta 12/0 ahead (CI/dependabot-only mirror); do NOT double-count. [APPLIED-METTA-BIOMEDICAL-NL-KG-CONSUMER-STACK] / [MIXED-COUPLING-RUNTIME-CONSUMER-PLUS-TEXT-EMITTERS] / [ICOG-CANONICAL-ORIGIN-REJUVE-DOWNSTREAM].d128ba91, active; [SUBREP-CERTIFIED-SKILL-DISCOVERY-HYPERON-CERTIFICATE-STORE]) — the program's first implementation-backed SubRep: a Python RL skill-discovery loop (2-head payoff/motive MLP over MO-LunarLander) with code-backed CDS/PDS admission gates (cite SubRep paper Sec 3.2 Def 1/2) and a REAL Hyperon certificate store (hyperon>=0.2.0; certificate↔MeTTa-Atom bridge, add/query/serialize, round-trip tested). Coupling SCOPED to certificate storage/query (not a full skill-execution runtime). Paper-aligned, NOT AIRIS-derived — closes the HAA S1 [SUBREP-AIRIS-EXTENSION-NOT-SOURCE-SPEC] narrative guard. Full write-up: Publication Maps+SubRep Deep Dive. Qwestor-PeTTa (@7f2b701c; [PETTA-LLM-CONTEXT-ACTION-ROUTER]) — a PeTTa MetaMo-style LLM-context-driven action router (Gemini-2.5-Flash context parse → appraisal/homeostasis/goal-weighting/routing-guards → action selection); pure PeTTa source + Python helpers, no in-repo hyperon runtime; NOT a query/KG system. Pattern-miner-mm2 (@d95b577c; [MM2-SURPRISINGNESS-MINER-PARTIAL-IMPLEMENTATION]) — a tiny MM2/MORK-native surprisingness miner (surp.metta validated vs hyperon-miner isurp-old: 3 pass / 5 blocked on 4-conjunct + N-ary gaps; frequent-miner.metta is an empty stub); a NEW MM2-native experiment, distinct from opencog/miner + hyperon-miner (HAA S3) — do not double-count. PeTTa-OpenPSI (@f59d3c5c; [PETTA-HYPERON-OPENPSI-PORT-WITH-PLN-NARS-USECASES]) — a PeTTa/Hyperon port of the OpenPsi motivation/emotion/action loop (demands/modulators/feeling-updaters) + a metta-trader use-case adding PLN (deduction/conjunction wired; abduction/induction disabled) + NARS (NAL-1); MIXED coupling (hyperon==0.2.3 + Python @register_atoms hooks; borrows TruthValue from hyperon-pln), NOT pure MeTTa. THIRD OpenPsi artifact — the PeTTa-stack successor; cross-link to the OpenPsi pilot (OpenCog-era) + hyperon-openpsi (HAA S2 Python predecessor), do not merge. [ICOG-NATIVE-MIXED-RESEARCH-LINES-SUBREP-IMPLEMENTATION-BACKED].c85f077f, origin/dev; [METTA-RAG-DEV-ASSISTANT-WITH-FRONTEND-HYPERON-WASM-PLAYGROUND]) — the IN-SCOPE standout: a MeTTa/Hyperon developer assistant whose backend is RAG over MeTTa code/docs (in-house AST chunker, SentenceTransformer embeddings, Qdrant, FastAPI; no hyperon runtime) and whose frontend ships a real Hyperon v0.2.10 WASM playground (Rust→WASM run_metta via hyperon::metta::runner::Metta). MIXED coupling (text-RAG backend + in-browser Hyperon runtime); active (dev 242/0 ahead of stale main). mcp-xp (@b63d7d49; [MCP-GALAXY-MULTI-LLM-BIOBLEND-TOOL-BRIDGE-ZERO-HYPERON]) — a zero-Hyperon Galaxy MCP bridge: both an MCP client (Azure/Groq/Gemini chatbot) AND a FastMCP server (bioblend_server exposing Galaxy tools + a galaxy://whoami resource) wrapping the BioBlend API, plus a Galaxy workflow executor. Absorb-value = the custom-MCP-server + multi-LLM-client pattern (aligns with the team MCP architecture + HERMES SC-3 delegation). galaxy-agent-xp-II (@086c5c52; [GALAXY-LLM-EMBEDDING-RECOMMENDER-ZERO-HYPERON]) — a zero-Hyperon FastAPI Galaxy tool/workflow recommender (SBERT/E5 cosine over Galaxy + IWC-GitHub-workflow metadata + OpenAI/Gemini classification); Galaxy-bioinformatics-specific (boundary). [ICOG-AUGMENTATION-TECH-METTA-RAG-MCP-GALAXY-TOOLS].c210bbea, active, tag core@0.4.7; [DECENTRALIZED-UAGENTS-FRAMEWORK-WITH-ALMANAC-BOUNDARY]) — a decorator-based async agent framework: Agent runtime + on_message/on_interval handlers + typed Protocol/dialogue state machines + crypto identity + Agentverse mailbox transport. Its decentralized USP is the Almanac: agents register/resolve via a CosmWasm smart contract on the Fetch.ai ledger (cosmpy; API-preferred + ledger fallback) — the only blockchain touchpoint, narrow (registration/discovery, not a cognitive chain layer). An experimental ChatAgent runs ASI:One (asi1-mini via litellm) with tool use, and a uagents-adapter ships LangChain / CrewAI / MCP / A2A adapters (code-backed MCPServerAdapter). fetchai/agentverse-skills (@16fab531; [PORTABLE-SKILLMD-AGENTVERSE-WORKFLOW]) — the first SKILL.md-format skills for the Fetch.ai/ASI ecosystem (anthropics/skill-md-spec — the SAME open standard as HERMES SC-3): 8 SKILL.md + 8 runnable scripts (chat/deploy/image-gen/inspect/manage/memory/search/asi1-chat) consumable by AI coding agents (Claude Code/Codex/Cursor/Gemini CLI) via a shared relay that deploys a temp hosted Agentverse agent; asi1-chat = ASI:One LLM, agentverse-memory = an Agentverse Memory MCP service. Ecosystem map: fetchai SDK [legacy, superseded by uagents-core] + ai-engine-sdk-python [AI Engine function-execution] + langchain-uagents [LangChain→uAgent adapter] + AEVS-sdk [tamper-evident HMAC/hash-chain tool-call receipts; LangChain/MCP interception]. Boundary: all S1 = zero-Hyperon augmentation tech; blockchain/ledger/wallet/token (cosmpy/fetchd/agent-launch-toolkit tokenization) excluded; ASI:One/AI-Engine/Agentverse shared with SC-1 asi-alliance. [EXTERNAL-AGENT-FRAMEWORK-AND-SKILL-WORKFLOW-TO-ABSORB]. Full workflow write-up: External Agent Workflows+uAgents Agent Framework.71f35a5b) — example-scoped, NOT a runtime contribution. Three official uAgent chat examples front an IN-PROCESS Hyperon MeTTa knowledge space (from hyperon import MeTTa + MeTTa() + native atoms via metta.space().add_atom(E(S(…))) + metta.run(); hyperon>=0.2.6 + uagents>=0.22.5), each a uAgent + ASI:One LLM (intent/wording) over a MeTTa structured-retrieval layer: web3/singularity-net-metta (Medical RAG; [MEDICAL-UAGENT-HYPERON-METTA-RAG-BRIDGE]), web3/singularity-net-metta-fetch.ai (Fetch/uAgents FAQ RAG; [FETCH-UAGENTS-FAQ-HYPERON-METTA-RAG-BRIDGE]), web3/singularity-net-metta-fetch-financial-advisor-agent (Investment RAG; [FINANCIAL-UAGENT-HYPERON-METTA-RAG-BRIDGE]). The broader A2A (a2a-uAgents-Integration Multiagent-Planner, a2a-cart-store; [UAGENTS-A2A-INTEGRATION-EXAMPLES-ZERO-HYPERON]) and MCP (mcp-agents/ticketlens, anthropic-quickstart/04-mcp-agent; [UAGENTS-MCP-INTEGRATION-EXAMPLES-ZERO-HYPERON]) examples are zero-Hyperon integration patterns. Coupling is narrow: all 13 hyperon hits are inside the 3 bridge dirs, 0 outside; NO AtomSpace/PLN/MORK/PeTTa (MeTTa-space demos, not those runtimes). [UAGENTS-HYPERON-METTA-BRIDGE-EXAMPLES-PLUS-A2A-MCP-INTEGRATION].[FETCHAI-AGENT-FRAMEWORKS-WITH-NARROW-HYPERON-BRIDGE-EXAMPLES]. Fetch.ai is the ASI-Alliance's decentralized agent-network infrastructure + applied coupler — S1 = external agent-framework infra (uAgents decorator framework + Almanac on-chain discovery + agentverse-skills SKILL.md + MCP/A2A/LangChain adapters), zero-Hyperon augmentation tech to absorb (same anthropics/skill-md-spec standard as HERMES SC-3); S2 = applied coupling (a uAgent fronts an in-process native-atom Hyperon MeTTa space + ASI:One). Fetch.ai does NOT author MeTTa runtime — it provides the decentralized transport/execution pipeline that lets Alliance members (SNET/Hyperon) expose cognitive runtimes. The MCP/A2A integration thread culminates here: HERMES (SC-3) → mcp-xp (SC-6) → uAgents/A2A/MCP (SC-2). Workflow card: External Agent Workflows+uAgents Agent Framework.c82268ec; [AGENT-LTM-CHROMADB-IMPORT-UTILITY-METTA-AS-DATA]) — an agent long-term-memory ingestion utility: distilled knowledge (JSONL + curriculum.metta, read as text) → OpenAI/SentenceTransformer vector embeddings → a ChromaDB “memories” collection (feeds the Max/Oma agents' LTM at /PeTTa/chroma_db). Zero Hyperon runtime (imports chromadb/openai/torch, no hyperon dep) — a clean closing illustration of the program's coupling discipline: the packaged KB data is full of Hyperon/MeTTa/MORK mentions, but the executable importer never runs Hyperon (knowledge-content ≠ code coupling). Oma_folio + Max_folio ([CLAW-AGENT-OUTPUT-FOLIOS-SURVEY-ONLY]) — survey-level artifact/output repos of the SC-7 OmegaClaw/MeTTaClaw agents in production (Oma the crab: 7200+ cycles, MeTTa/NAL/PLN reasoning, sonifications; Max Botnick: NAL-forward-chaining certification reports, hypothesis papers, MeTTa artifacts) — maturity evidence for the SC-7 Claw line, NOT new framework code (cross-ref OmegaClaw-Core → SC-7). Boundary: ASI-Chain (10 repos = F1r3fly blockchain; asi-chain-node → SC-4) + max-music (generative-music album) excluded; ASI:One/AI-Engine/Agentverse → SC-2. [ASI-THIN-CLOSE-AGENT-LTM-UTILITY-PLUS-CLAW-ARTIFACTS].trueagi-io/FormalMeTTa is HTTP 404, corrected 2026-05-15). Partial: library-only, carries unimplemented ??? holes (e.g. RewriteRules.scala:136). Archival.[CLASSICAL-ATOMSPACE-BRIDGE-STUB] / [ASMETTA-FUNCTION-DECL-IMPORT-ONLY] / [ASMETTA-CORE-SEMANTICS-UNIMPLEMENTED-CONFIRMED]. On the AtomSpace-Scheme side of the Perception two-stack split. Archival 2023 (S6 closed 2026-06-08).Inference engines, learning algorithms, and autonomous agent reasoning systems.
[GNN-NEURAL-MINING] / [PARALLEL-NON-INTEGRATED] with hyperon-miner-family symbolic miners. See Neural Pattern Mining. Active.an_infotheoretic_approach/ (info-theoretic with GA + CMA-ES), peTTa_version/ (PeTTa port with multi-objective CMA-ES), a_categorytheoretic_approach/ (category-theoretic colimit/pushout), a_naive_approach/ (graph matching + GPT prototype). Per Non-clustered HAA cluster pilot Source 4 close 2026-05-07: [CONCEPT-BLENDING-PARTIAL-IMPLEMENTATION] — info-theoretic + GA/CMA + category-theoretic colimit branches are [IMPLEMENTATION-BACKED-CORE] at HEAD 9675e5e; quantale + paraconsistent + HoTT-adjacent material is [PAPER-ONLY-FORMAL-EXTENSION]. See Concept Blending. Active.[ACTIVE-METTA-PRIMARY] at HEAD 9f69e17 (2026-03-26); the canonical Hyperon-era MOSES line. See MOSES. Active.[ACTIVE-CPP-ATOMSPACE-BASELINE] at HEAD dde4849c4 (2025-12-10). linas/as-moses is byte-identical at the same SHA [VEPSTAS-MIRROR-SAME-SHA] — not a separate implementation tradition. Maintained.[SUPERSEDED-BY-METTA-MOSES] for general MOSES; retained as boolean-reduction / ENF-pipeline prototype. P12-verified 27 commits ahead of ngeiswei/hyperon-moses older sibling. Active.[ARCHIVAL-PARADIGM-PREDECESSOR] — last commit 2023-09-19. Archival.[STRICT-FORK-STALE] (0 ahead / 107 behind); singnet/asmoses [STRICT-FORK-STALE-MERGED-UPSTREAM] (7 ahead = merge-sync metadata / 165 behind); singnet/moses [STRICT-FORK-DIVERGED-LEGACY-FIXES] (19 ahead / 23 behind). None contribute new traditions. iCog-Labs-Dev/moses-optimization-algorithms is [STANDALONE-EXPERIMENT-NOT-FORK] — independent commit history (Initial commit / hill climbing / simulated annealing / Nahom32 PR), not derived from asmoses or moses.[OPENPSI-PREDECESSOR-SUBSTRATE] — at HEAD 3b356c5 contains zero MetaMo/MAGUS naming, so this is NOT a MetaMo strict-literal implementation; the appraisal comonad in MetaMo abstracts this substrate at one level higher. See MetaMo. Active.[HEURISTIC-PROTOTYPE] — no formal categorical structure; targeted searches at HEAD for "bimonad" / "comonad" / "MAGUS" / "OpenPsi" return zero matches. Parallel-paradigm to MetaMo-Python. Active.[IMPLEMENTATION-BACKED-NAL1-5] / [METTA-NARS-NOT-PLN] at HEAD 945e2f8 (2024-10-14); ONA-style NARS port with f/c truth values. Paradigm-distinct from PLN — cross-link as boundary, not as alternative implementation. See MeTTa-NARS. Active.[IMPLEMENTATION-BACKED-CORE] / [AIRIS-DERIVED-NAL-ADJACENT] / [NACE-NOT-AIRIS] at HEAD 361fddf (2025-05-06); pure-Python (no src/ or C++; PyQt GUI in gui.py); extends AIRIS apparatus with partial observability + NAL evidence; bridges to MeTTa-NARS via bridge.py. The "1000x data efficiency" framing should be paper-sourced if retained. See NACE. Active.singnet/AIRIS_Public (HTTP 404); URL corrected at cluster-close. Active.[ACTIVE-IDRIS-DSL] / [AI-DSL-DUAL-CITIZEN] at HEAD aaa702e (2025-06-25); two implementation tracks — Idris dependent-types layer (77 .idr files in experimental/ai-algorithms/) and MeTTa combinator/ontology layer (30+ .metta files in experimental/ai-service-composition/). Center of gravity is service composition for the SingularityNET marketplace; the MeTTa combinator track is the HAA-relevant slice. NOT a core cognitive algorithm comparable to PLN/ECAN/MOSES/NARS. See AI-DSL. Active.singnet/semantic-parsing here returned HTTP 404 and was corrected to this trilateral set):
[SEMANTIC-PARSING-SPLIT-STACK] — these are LLM-to-PLN/MeTTa, NOT wrappers around classical Link Grammar / RelEx (which see under Natural Language Processing). See Semantic Parsing.
Hypergraph databases, storage backends, and distributed data infrastructure.
[PATTERN-INDEX-SEPARATE-PARADIGM-ADJACENT] — adjacent legacy infrastructure to Pattern Mining, distinguished by RAM-only AtomSpace mining (the symbolic paradigm) vs disk-based subgraph indexing (this repo). Maintained / Archival.On-chain execution environments and decentralized AI infrastructure.
Domain-specific applications of Hyperon capabilities in longevity research, bioinformatics, and enterprise AI.
Agent frameworks, virtual environments, and embodied AI testbeds.
[SCAFFOLDING-NOT-PIPELINE]. See Vision. Scaffolding.Tools, training materials, IDE support, and integration utilities for MeTTa developers.
[IMPLEMENTATION-BACKED-CORE] for the X = G × M skeleton + F = D ∘ Ψ pseudo-bimonad; [FORMAL-LAWS-PAPER-ONLY] for categorical proofs; primary formal reference for MetaMo (over hyperon-openpsi which is the OpenPsi predecessor substrate, and MetaMo-Prototype which is heuristic). See MetaMo.[DEMO-NOT-PIPELINE].Core C++ framework components. Maintained for AtomSpace compatibility; not the Hyperon-forward path.
Related cards: Implementation Families · Technical Companions · TrueAGI · iCog Labs