Skip to content

Engraphy v0.1.0

Choose a tag to compare

View all tags
@devon-clarkk devon-clarkk released this 22 Aug 14:29
· 27 commits to main since this release
v0.1.0
3465fc8

Engraphy is a self-hosted memory engine for AI agents. It stores what an
agent learns as a typed knowledge graph on Postgres + pgvector, and speaks the
Model Context Protocol, so any MCP client (a VS Code extension, a desktop app,
another agent) can use it over HTTP.

It exists to replace the reference MCP memory server's flat-JSON, single-user,
stdio model with something that survives concurrency, paraphrase, duplicates,
and years of accumulated memory.

What is in this release

  • A typed memory graph. Memories are typed nodes (fact, decision,
    person, event) joined by typed edges (involves, references,
    supersedes). The types, their attribute schemas, and the rules for which
    edges may connect which types are declared per space in a pack and
    enforced in Postgres.
  • Writes that deduplicate themselves. Every write is embedded and banded
    against existing memory: a near-verbatim restatement auto-merges, a genuinely
    new but related fact is kept and linked, and a borderline case parks as a
    pending verdict for the caller to resolve. Nothing is silently absorbed.
  • Hybrid retrieval. search fuses a vector leg and a Postgres full-text leg
    with Reciprocal Rank Fusion; traverse walks the edges.
  • Isolation the database enforces. Spaces and principals are separated by
    Row-Level Security under a NOBYPASSRLS role, not by application checks that
    can be forgotten.
  • An operator CLI and an MCP tool surface, covering everything from
    bootstrapping a space to minting tokens, applying packs, and verifying
    restores.

Running it

Requires Docker with Compose. From a checkout of this tag:

cp deploy/.env.example .env   # then edit, or generate the two passwords:
printf 'POSTGRES_PASSWORD=%s\nENGRAPHY_APP_ROLE_PASSWORD=%s\n' \
  "$(openssl rand -hex 16)" "$(openssl rand -hex 16)" > .env

docker compose up -d          # first boot downloads the ~523 MB embedding model

docker compose --profile admin run --rm admin \
  engraphy-admin space create --id personal --display-name "My Memory" --principal me
docker compose --profile admin run --rm admin \
  engraphy-admin pack apply packs/starter/pack.yaml --space personal
docker compose --profile admin run --rm admin \
  engraphy-admin token create --space personal --principal me \
    --client-name my-editor --role readwrite

The MCP endpoint is then http://127.0.0.1:8000/mcp/ (keep the trailing
slash). Put a TLS-terminating reverse proxy in front before exposing it. The
bearer token prints once.

Note: up.sh / provision.sh wrap this sequence and wait for /healthz.
They landed after this tag, so use them from main rather than from the
v0.1.0 tarball.

Documentation

License

Source-available under the Business Source License 1.1. You may read,
modify, redistribute, self-host, and use Engraphy in production as the memory
layer for your own applications and agents. You may not offer Engraphy itself
to third parties as a hosted or managed service before the Change Date
(2030-08-22), when it converts to Apache-2.0. See
LICENSE.

Status

v0.1.0, one maintainer. The schema and enforcement kernel, the engine
behaviours (dedup, hybrid retrieval, traversal), the pack system, and the
operator CLI are all in the shipped engine.

Assets 2
Loading