mushroomdb

The graph that stays true — and knows who's allowed to see it.

mushroomdb is the data layer for agents that reason over entities. It is an embedded Rust graph database in which a relationship is a schema declaration: write a rule once, and every write derives, maintains and retracts the matching edges, each one carrying the rule, the score and the values that produced it. An agent reaches it over MCP — sixteen tools on an entity store — or you embed it as a Rust library, a Python module, or a sidecar beside your own service. Four questions are what it exists for: why are these two related (explain_association, answered with the evidence rather than an assertion), what did that look like then (edges_at, the edges a node had at any past commit), what would this change do (what_if, computed without writing anything), and who may see it (query with a role, so one graph answers differently per caller). Local-first: a directory on disk, no account, no endpoint, no model call in the write path unless you enable embeddings.

ingest-git stays supported as a data source: commits, pull requests, files and authors become entities with rule-derived relationships, which is what makes a ticket↔commit link a rule rather than a script.

Deprecated in 0.6.4, removed in 0.7: the code-graph door — seven tools, three grep/edit hooks, and the plugin's coding-assistant positioning. Still working, still tested. What this means for you.

Pre-1.0 alpha — APIs and formats may change between minor versions.

Docs · Changelog · Issues

Quick start

npx mushroomdb install --db ./memory

One command writes the /mushroom skill, an MCP server listing the sixteen-tool association surface, and the session hooks. Then a worked flow, four tool calls:

upsert_entity  →  create_rule  →  find_similar  →  explain_association
  (store)           (link)           (recall)          (explain)

Full tool reference: docs/site/mcp.md.

• Live, not a snapshot. One SET on a property re-derives the matching edges — added, scored and retracted — before the write closes. The SET in the GIF above is that one write.

• Retracts instead of going stale. A property drifting out of a rule's predicate doesn't leave a stale edge behind; the engine retracts it in the same write that caused the drift.

• Explains any link. explain_association names the rule, the score, and the values the two entities actually share, so "why are these two related?" has an answer your assistant can quote instead of a guess.

• Knows who's allowed to see it. Pass a role or a mask with a query and the same graph answers differently per caller; write statements are rejected on masked queries.

• Answers what it said last week. edges_at returns the edges a node had at a past commit; mushroomdb asof ./db --commit 5 --query "…" replays the WAL to that commit, derived edges included.

Related MCP server: Graphiti Memory MCP Server

Where it fits

What it is

• An embedded, single-binary graph database with a rule engine that maintains edges for you.

• A 28-tool MCP server — sixteen listed on an entity store, three on a store built by ingest-git — plus a /mushroom skill and a Claude Code plugin.

• Safe for several processes at once: one writer at a time behind an advisory LOCK file, any number of readers, and every handle picks up a peer's commits by refresh() rather than reopening — so a running serve, an editor hook, a git hook and a CLI command can share one store. docs/site/concurrency.md

• Local-first: your data stays on disk, no cloud service, no model call in the write path unless you enable embeddings.

What it isn't

• Not a hosted memory service — there is no account, no endpoint, nothing to sign up for.

• Not a vector database. Vector predicates and HNSW are built in; bring your own embeddings.

• Not a Postgres replacement. Single writer, no interactive transactions, memory-first storage.

The differentiator

Most graph databases require you to create edges manually or run a batch similarity script after each load. mushroomdb makes edge creation a schema declaration. A rule like "connect every Person to every Org whose skills list overlaps theirs by at least 50%" is written once:

db.create_rule(RuleDef {
    name: "skill_fit".into(),
    src_label: "Person".into(),
    dst_label: "Org".into(),
    predicate: Predicate::Overlap { field: "skills".into(), min: 0.5 },
    edge_type: "FIT".into(),
    weight_prop: Some("score".into()),
    max_edges: Some(5), // keep the 5 best-matching Orgs per Person (top-k per source)
}).expect("rule");

After that, every insert_node and set_prop evaluates the rule incrementally. The engine writes the edge, stores the Jaccard score, and retracts the edge if the properties later diverge — without any manual work.

Watch it live — a Cypher SET changes one property, the founded_within rule fires, and new scored edges appear in the bundled explorer:

Open the Rules panel, and the Why slide-over shows the exact predicate arithmetic behind every derived edge:

Predicates

Six predicate kinds ship today. They compose via All(...) (AND, score = min) and Any(...) (OR, score = max), nested up to depth 4.

Auto-FK: fields ending in _id whose values match existing node keys get KeyMatch rules created automatically at ingest time. VectorSimilar accepts approximate: true to switch candidate selection to in-tree HNSW (per-query recall floors min 0.90 / mean 0.95, measured 1.0 / 1.0 at 5k nodes / dim 1536, fixed-seed probe). Full reference: docs/site/rules.md.

Built on the same engine

• Live subscriptions. subscribe_rule (Rust) and GET /subscribe (WebSocket) stream EdgeFired / EdgeRetracted the moment they hit the WAL — not polled, not batched. Bounded 65,536-event queue; slow consumers get a Lagged { missed: N } marker instead of a disconnect. docs/site/subscriptions.md

• Rule attribution across time. Every derived edge writes a HISTORY-MARKER WAL record carrying the rule name, so edge_history, node_history, and was_linked answer which rule created a link and at which commit. GraphDb::open_at(&dir, 5) replays to a past commit, derived edges included; out-of-range commits return CommitOutOfRange, never wrong data. docs/site/timetravel.md

• Materialized views. Degree counts and neighbor aggregates (sum/avg/min/max) maintained incrementally on every edge change — no cron, no triggers, no stale caches. docs/site/views.md

• Rule suggestions. db.suggest_rules() (or mushroomdb suggest ./db) profiles your data and ranks candidate rules with estimated edge counts and rationale. Seeded sampling, so the same database always returns the same suggestions. No rule is ever applied automatically. docs/site/suggest.md

• An exception class per engine error. In Python, MushroomError(RuntimeError) is the base and eighteen subclasses hang off it — KeyNotFound, ReadOnly, Corrupt, CasConflict, RoleWriteDenied and the rest — each carrying .code, a stable snake_case string, and the variant's own fields as attributes. .code is the compatibility surface: classes may be added, a code is never respelled. Because the base subclasses RuntimeError, every except RuntimeError written against an earlier release keeps catching what it caught. A test asserts every Rust GraphError variant maps to a distinct class, so adding one without a class fails the build. db.roles() reads back what roles.json defines, so a sidecar can validate a role name at boot rather than on the first request.

Deprecations

What "deprecated" means here. It still works in 0.6.4, it is still tested on every release, and nothing is removed. It is no longer promoted — not on this page, not in the skill's task rules, not in the plugin's description — its documentation page opens with a notice, and an install that turns one of the hooks on prints a deprecation line. It is removed in 0.7. The migration is nothing to do, unless you relied on the specific thing named below.

Why, measured. Across 240 cells — arms stock / installed / invoked / cli × 3 reps × 20 tasks over two repositories — the invoked graph arm scored 0.924 against stock's 0.927 (paired difference -0.0035 [-0.0112, 0.0017]) and cost $0.2713 against $0.2267 (+0.04463 [+0.01312, +0.07471], about +20%), and the two installed-but-not-invoked arms made 0 graph calls in 120 sessions: results/20260910T000418Z. An agent holding grep neither needs a graph for those questions nor chooses one. What the engine is for is measured separately, on the association suite, and reported under Benchmarks.

Agent memory

Graph structure captures the shape of real knowledge — entities, associations, similarity, and lineage — and rule-derived edges keep those associations fresh as new facts arrive.

• Entities map to nodes (Person, Document, Project, Concept, …).

• Associations are edges derived from data: cosine similarity on embeddings, shared field values, FK relationships, geographic proximity. Declare a rule once; every write maintains the matching edges without agent-side bookkeeping.

• Recall has three modes: find_similar by query vector (HNSW when available, brute force otherwise); find_similar by key (neighbors along a rule-derived edge type); query for structured Cypher recall. hybrid_search fuses fulltext and vector results via Reciprocal Rank Fusion, and pairwise_similar answers "which of these are most like each other" over a caller's own key set, exactly — it never uses HNSW.

• Explanations are built in: explain_association shows which rules and scores produced each link, so an agent can cite evidence instead of asserting a conclusion.

• Scoping is a property of the handle, not of each call: db.scoped(role=…, namespace=…, keys=[…]) returns a read-only child sharing the same store, and every read on it obeys one contract — hidden behaves exactly as absent, so a key outside the scope is indistinguishable from a key that does not exist. Legs intersect, so a scope narrows and never widens; every mutation raises ReadOnly. Per-call mask: [key1, key2, …] on query still works and still rejects writes. Both are cooperative in-process argument handling rather than an access boundary — real enforcement is the HTTP server's role tokens. docs/site/masks.md

• Bulk loading is one atomic frame: ingest_batch(nodes, edges, on_conflict="error" | "skip" | "replace") lets a mirror rebuild onto a store that already has content without wiping the directory, and reports inserted, edges_inserted, skipped, replaced and kept_view_owned. "error" is the default and is the older behaviour exactly.

• Schema-as-code: mushroomdb schema apply <dir> <schema.json> idempotently applies rules, views, and fulltext indexes, printing a created/updated/unchanged diff.

Minimal workflow (four tool calls):

upsert_entity  →  create_rule  →  find_similar  →  explain_association
  (store)           (link)           (recall)          (explain)

Fourteen task tools answer a question in prose in one call. Seven answer on any store; seven are the deprecated code door. They are what the skill reaches for, and what tools/list shows first:

Each of the fourteen also takes json: true, which answers with the raw report instead of the rendered digest.

The fourteen graph tools reach the store directly. Their descriptions are prefixed Advanced: in tools/list, so an assistant knows which surface is the front door. The default listing follows the store: a store built by ingest-git lists three tools in all — explore, query and stats — and any other store lists sixteen, the association surface: query (with an optional role), explain_association, neighborhood, node_info, node_edges, was_linked, edges_at, what_if, node_history, edge_history, find_similar, pairwise_similar, hybrid_search, remember, recall and stats. All 28 stay served either way — the listing decides what a session can call, not what the server answers — and mushroomdb mcp <db> --all-tools lists the whole set:

Full walkthrough, tool reference, and Claude Desktop setup: docs/site/mcp.md. Skill, plugin, and hook details: docs/site/skill.md.

Install options

claude plugin install mushroom@mushroomdb   # after `claude marketplace add MatthewSherlin/mushroomdb`
npx mushroomdb install            # skill + MCP server + hooks, no toolchain needed
cargo install mushroomdb-cli      # `mushroomdb` binary from crates.io (no embedded UI)
cargo add mushroomdb              # embedded Rust library
pip install mushroomdb            # Python bindings

install writes an MCP entry that runs npx -y mushroomdb@<version>, so the assistant needs nothing installed globally and nothing is copied into your home directory. Point it at a local build with --command <path>. mushroomdb doctor verifies the result end to end — config entry, store, lock, hooks, git hooks, and a real stdio handshake with the configured command.

To see the bundled explorer, write a demo graph and serve it:

mushroomdb demo ./db
mushroomdb serve ./db

Open http://127.0.0.1:8080/. The demo graph has 10 Orgs, 20 Projects, 30 People, and 334 edges — 304 of them derived by seven rule sets. When a token is configured, open http://host:8080/?token=…. Building the binary with the UI embedded, Docker, and the install.sh script are covered in CONTRIBUTING.md.

Role-bound tokens limit a caller to a named subset of nodes. Define roles in schema.json under the roles key (each role has a label selector list), then pass --role-token TOKEN:ROLE (repeatable) when starting the server, or set MUSHROOMDB_ROLE_TOKENS="tok1:role1,tok2:role2". A role token receives only the nodes matching its label selectors — read endpoints return rows filtered to the visible set; write, subscription, and analytics endpoints return 403. Unknown token or role name: 401. The never-widen invariant is enforced in the server: a client-supplied mask is always intersected with the role mask. The MCP interface (mushroomdb mcp) is a stdio JSON-RPC server for local agent use and is not subject to bearer-token or role enforcement.

CLI reference

Concurrency: every CLI write command, the hooks, and a running mushroomdb serve coordinate through one advisory LOCK file in the store directory, so they are safe to run against the same store at the same time. A writer that cannot get the lock within two seconds exits 3 with another mushroomdb process is writing; retry, having written nothing. Readers never take the lock and never wait; recall opens read-only (read_only: true) so an unattended hook can never delay a writer or fail because one is running. What the lock does not give you: cross-process transactions, and subscription events for a peer's writes — a commit absorbed by refresh() is visible on the next read but notifies nobody. Full model: docs/site/concurrency.md.

Full HTTP endpoint reference: docs/site/api.md.

Known limitations

Benchmarks

10,000-node graph (Apple M4 Pro, macOS 15.7.3, arm64), mushroomdb v0.1.1 release build, 2026-08-24. Full methodology and honesty notes: benchmarks/results/head-to-head-10k-v2.md.

Honesty notes:

• mushroomdb numbers are embedded — no network round-trip, no serialization overhead. KùzuDB is also embedded, so its numbers are directly comparable. Neo4j and Memgraph go over bolt/localhost (~0.1–1 ms round-trip per query).

• ★ Two-hop join: same dataset, same warmup policy, all four engines on 5,810,000 INDUSTRY_ALIGNMENT edges. Fresh process → ingest + preload → 3 discarded warmups → median of 10 runs. mushroomdb derives the edges via create_rule; competitors were pre-loaded via UNWIND MERGE or COPY FROM CSV. All engines return 200 rows.

• ★ Earlier v2.1 two-hop values were retracted for cross-engine contamination; the v2 mushroomdb 307 µs figure was retired (measured on a smaller 1M-edge graph). Both are documented in the methodology file rather than quietly dropped.

• ▽ 100k cold-start measured 2026-08-28, warm file cache, cold process, /usr/bin/time -l: V8 snapshot open 0.02 s at 31–41 MiB RSS; snapshot size 1.8 GiB; snapshot write ~35 s. Cold-cache was not measured. See dogfood/results/scale-100k.md.

• Rule engine vs hand-rolled maintenance (10k nodes, 1,000 specialty updates, drift = 0 for all three): per-op expert-written 64.93 min, batched expert-written 24.98 s, rule engine 17.58 s. Both hand-rolled variants were written by the engine team with full knowledge of retraction semantics — drift = 0 is a property of that, not of hand-rolling in general. benchmarks/results/handrolled-vs-rules.md

Agent benchmarks measure the entity engine directly. benchmarks/agent-tasks/ runs real claude -p sessions against executable truth. The association suite (--suite association) asks twenty relationship questions of one generated world written three ways — as JSON files, as a single relational file, and as a mushroomdb store. The graph arm moved from 0.795 at $0.5089 (20260911T005749Z) to 0.967 at $0.0923 (20260911T065400Z) against the relational baseline's 0.974 at $0.1158 — a correctness tie at a lower mean cost, and still behind files + grep (1.000 at $0.0949) on a 2,000-entity world; the gate failed on the cost interval. The code suite (--suite code) is retired; its committed summaries stay as the record. docs/site/association-bench.md describes the suite and how to rebuild the world.

Architecture

graph-db/
├── crates/
│   ├── core-storage      # Packed adjacency topology + columnar property store + WAL + snapshots
│   ├── core-rules        # linking rules, per-rule indexes, incremental maintenance
│   ├── core-query        # pull-based interpreter; traversal ops + Cypher subset
│   ├── core-api          # the one public Rust interface; typed error enums
│   ├── code-extract      # tree-sitter symbol/import/call extraction; bytes in, facts out
│   ├── arrow-bridge      # results ↔ Arrow buffers
│   ├── server            # axum HTTP + WebSocket; serves UI
│   ├── cli               # mushroomdb binary
│   └── sim-harness       # DST: virtual clock, fault-injecting IO, seeded runner
├── ui/                   # TypeScript + Vite graph explorer
├── bindings/python/      # PyO3 / maturin
└── clients/typescript/   # HTTP + WebSocket client

Dependency rule (inward only): bindings/server/cli → core-api → {core-query, core-rules} → core-storage

Storage uses a dense-id WAL with per-commit fsync (configurable via FsyncPolicy), plus mmap-able V9 rkyv snapshots (13 sections: CSR topology, columnar properties, one shared string table for every string column, HNSW blobs, provenance, IVF state, per-node last-change index, and more — zero-copy, no heap allocation on open). V5–V8 stores are auto-migrated to V9 on GraphDb::open, keeping the original beside the new one as snapshot.bin.bak until the next clean open. The upgrade is one-way: an earlier binary refuses a V9 snapshot with snapshot: unsupported version 9 rather than reading it wrongly. Derived edges are not WAL-logged; they are restored directly from the mmap'd sections. See docs/format-stability.md for the format evolution contract.

Roadmap

Phases 1–4 and Plan 18 all landed. What remains:

Docs

• Quickstart · Rules · Cypher reference · HTTP + MCP API

• The live code graph (deprecated in 0.6.4, removed in 0.7) · Concurrency · Codebase graph

• Install, plugin and hooks · MCP tools · Association benchmark

• Time travel · Subscriptions · Views · Rule suggestions

• Masks and access control · Full-text search · Property indexes · Graph algorithms

• Durability and recovery · Running it as a service · Panic policy · Testing · Format stability

• Design spec · Moat roadmap · Case study

Building from source, Docker, packaging, and the test gates are in CONTRIBUTING.md.

License

Copyright 2026 Matthew Sherlin.

Dual-licensed under MIT or Apache-2.0, at your option.