VSA and local LLM AI engine
  • Rust 98.4%
  • Python 0.8%
  • HTML 0.3%
  • CSS 0.2%
  • Shell 0.1%
  • Other 0.1%
Find a file
Till Wegmueller 13aead9a2e
Some checks failed
CI / publish-chart (push) Failing after 1s
CI / docker-seshd (push) Failing after 14m7s
CI / check-seshat (push) Failing after 14m22s
docs: scaffold-merge audit — 35 hollow .anima-only merges classified
6 genuine phantoms (Phase 38 #65, AGE image #67, OCI models #63, Akhipedia
drawer-extraction #122, heka bridge #123, home-state fluents #152) + #162/#163
reopened. Rest backed by real code via sibling PRs. Full per-PR table.
2026-06-02 17:29:22 +02:00
.anima/wip Merge PR #163: wip: triage Entity:Relation ratio 4.95:1 — relation symbol deficiency after cleanup 2026-06-02 14:41:46 +02:00
.cargo fix: Add C++ include path workaround for macOS 26 CLT 2026-02-16 19:02:32 +01:00
.claude-plugin fix: Idempotent triple insertion + Claude Code plugin scaffold 2026-03-18 22:16:50 +01:00
.forgejo feat: CI: docker-seshd + publish-chart marked failure even when artifacts upload (019dc4de) 2026-06-01 17:17:25 +00:00
.github/workflows feat: Add workspace hermeticity test to CI (019e0b78) 2026-06-01 17:01:48 +00:00
agents fix: Idempotent triple insertion + Claude Code plugin scaffold 2026-03-18 22:16:50 +01:00
benches Merge PR #165: feat: [VSA Agent] VSA-Native Action Selection in OODA Decide Phase (019e403b) 2026-06-02 12:58:11 +02:00
book docs: Fix binary names in book, add bootstrap chapter and systemd service 2026-03-02 11:00:52 +01:00
charts/akh-medu fix(seshat): migrate pgvector index from IVFFlat to HNSW 2026-04-25 00:32:34 +02:00
contrib/systemd fix: Filter bootstrap metadata from chat output, add ONNX Runtime to systemd 2026-03-02 13:20:25 +01:00
crates Merge PR #182: feat: [Gemma+Psyche] ELIZA + Gemma Engine Selector — route input to best NL handler (019e4084) 2026-06-02 13:53:25 +02:00
data Merge PR #213: feat: [HIGH] Ennead is hollow: 7 members exist as names only, not identities (019e587a) 2026-06-02 14:13:49 +02:00
docker ci: rewrite seshd Dockerfile for legacy builder (matches Anima CI pattern) 2026-04-21 22:23:33 +02:00
docs docs: scaffold-merge audit — 35 hollow .anima-only merges classified 2026-06-02 17:29:22 +02:00
hooks feat: [MEDIUM] Zero episodic memory: 7+ days of rituals never ingested (019e5356) 2026-06-01 16:01:23 +00:00
plugin-skills fix: Idempotent triple insertion + Claude Code plugin scaffold 2026-03-18 22:16:50 +01:00
scripts feat: [Gemma+Psyche] LoRA Fine-Tuning Pipeline for Psyche-Encoded Adapters (019e4084) 2026-06-01 16:32:59 +00:00
skills Merge PR #116: feat: Phase 30 — Procedural Skills (019db054) 2026-06-02 15:59:12 +02:00
src chore: clean up unused imports (cargo fix, 9 across 6 files) 2026-06-02 16:09:35 +02:00
tests feat: integrate PR #176 ELIZA chat routing classifier (reconciled into agent::eliza) 2026-06-02 15:49:40 +02:00
wit Add configurable tool registry with danger metadata, CLI tools, and WASM runtime (Phase 10) 2026-02-14 01:51:00 +01:00
.dockerignore feat: add Kubernetes deployment — Docker images, Helm chart, version 0.6.0 2026-04-14 22:53:58 +02:00
.gitignore chore: ignore .anima-exec/ runtime diagnostic dumps 2026-05-06 23:03:15 +02:00
.mcp.json chore: empty project .mcp.json (seshat moved to user scope) 2026-04-23 20:43:43 +02:00
AGENTS.md feat: [Gemma+Psyche] LoRA Fine-Tuning Pipeline for Psyche-Encoded Adapters (019e4084) 2026-06-01 16:32:59 +00:00
Cargo.lock chore: bump version to 0.7.0 2026-06-02 15:40:24 +02:00
Cargo.toml chore: bump version to 0.7.0 2026-06-02 15:40:24 +02:00
CLAUDE.md Merge PR #114: feat: Phase 24 — Sensory Grounding (019db054) 2026-06-02 14:57:25 +02:00
LICENSE Add GNU General Public License v3 2026-02-11 20:48:52 +01:00
README.md Update README.md 2026-04-22 21:30:33 +00:00
release.toml Use cargo-release for version management and tagging 2026-02-14 16:01:41 +01:00

akh-medu

A neuro-symbolic AI engine made to run on the CPU. Made for people. Not for robber barons.

Documentation

akh-medu combines hyperdimensional computing (Vector Symbolic Architecture) with knowledge graphs and symbolic reasoning. It runs entirely on the CPU with no LLM dependency, no GPU requirement, and no external NLP models.

  • Infer new knowledge via spreading activation, backward chaining, and superposition reasoning
  • Reason symbolically using e-graph rewrite rules (equality saturation)
  • Search semantically using 10,000-bit binary hypervectors
  • Act autonomously via an OODA-loop agent with 15 built-in tools
  • Parse and generate natural language in 5 languages via a grammar framework
  • Serve knowledge over REST and WebSocket APIs

Quick Start

cargo build --release
akh-medu init
akh-medu seed apply ontology
akh-medu grammar parse "Dogs are mammals" --ingest
akh-medu query --seeds Dog --depth 2

See the Quick Start Tutorial for a full walkthrough.

Architecture

┌──────────────────────────────────────────────────────────┐
│                      Engine API                          │
├──────────┬──────────┬──────────┬──────────┬──────────────┤
│   VSA    │Knowledge │Reasoning │Inference │   Agent      │
│  Ops     │  Graph   │  (egg)   │ Engine   │  OODA Loop   │
│ ─────────│──────────│──────────│──────────│──────────────│
│ HyperVec │petgraph  │ rewrite  │spreading │ 15 tools     │
│ SIMD     │oxigraph  │ rules    │backward  │ planning     │
│ ItemMem  │SPARQL    │ e-graphs │superpos. │ psyche       │
├──────────┴──────────┴──────────┴──────────┴──────────────┤
│                   Tiered Storage                         │
│  Hot (DashMap) · Warm (mmap) · Durable (redb)            │
└──────────────────────────────────────────────────────────┘

Documentation

The full documentation is hosted at akh-medu.dev and covers:

To build the docs locally:

cd book && mdbook serve

License

This repository is licensed under GPLv3. See the LICENSE file.

For integration into proprietary applications, contact the author.

Contributing

To make it possible to milk as much money as possible from proprietary vendors back into FLOSS work, every contributor must agree that we will sell very expensive proprietary licenses..